Autumn's here already! The long, hot summer that I had feared happened, I suppose, but it wasn't as bad as I feared. And somehow autumn is making itself more obvious than I would have expected; the horses are already changing coat, and the nights are getting cool.

Today the weather was pleasant, sunny and not too warm, so we decided to do a bit of tourism, and headed to the somewhat ineptly named Otway Fly, a tree-top walk through some of the Beech Forest part of the Great Otway National Park.

This isn't the first time I've been through this sort of thing. Six months ago we visited Maits Rest, another rain forest reserve not far away, and four years ago my father and I visited the Valley of the Giants in Western Australia, which was the first ever tree-top walk.

How do they compare? I found the admission price to “Otway Fly” to be surprisingly high, $19 (though we got $2.50 discount at the information centre in Colac). At the Valley of the Giants it was $6, and of course Mait's Rest was free.

But maybe it's more interesting? I'm still trying to make up my mind about that. I took lots of photos, so there's something to compare. From left to right, Giants Valley, Mait's Rest and Otway Fly:

Image title: Treetop walk 9          Dimensions:          1536 x 2048, 1222 kB Make a single page with this image Hide this image Make this image a thumbnail Make thumbnails of all images on this page Make this image small again Display small version of all images on this page All images taken on Thursday, 5 February 2004, thumbnails          All images taken on Thursday, 5 February 2004, small Diary entry for Thursday, 5 February 2004 Complete exposure details

 

Image title: otways 7          Dimensions:          2736 x 3676, 6389 kB Make a single page with this image Hide this image Make this image a thumbnail Make thumbnails of all images on this page Make this image small again Display small version of all images on this page All images taken on Wednesday, 10 October 2007, thumbnails          All images taken on Wednesday, 10 October 2007, small Diary entry for Wednesday, 10 October 2007 Complete exposure details

 

Image title: treetop walk 14          Dimensions:          2736 x 3694, 5883 kB Make a single page with this image Hide this image Make this image a thumbnail Make thumbnails of all images on this page Make this image small again Display small version of all images on this page All images taken on Saturday, 1 March 2008, thumbnails          All images taken on Saturday, 1 March 2008, small Diary entry for Saturday, 1 March 2008 Complete exposure details

 

On the other hand, there were other things to look at, including the biggest tree ferns I have ever seen. I'd estimate that these ones are over 6 metres across:

Back home, noting that despite the advent of autumn, bush fires are not behind us:

It was Yvonne's birthday today, so we made a fondue de fromage, this time using some “raclette” cheese that I got from Campana's in town. We'd had the raclette before, and found it not very typical, but it was just what we needed in the fondue.

Spent most of the day in the garden, planting bulbs. I'm wondering if it wasn't too early; the info on the packages suggested planting from February to June, and some (the hyacinths) are supposed to bloom as early as June (though presumably not if only planted in June). But the weather's still pretty warm, and today it went over 30°. Still, most of these bulbs can be left in the ground year round, so it's probably not very critical when you plant them.

A friend of Yvonne's, Barbara Sadler from Toronto, arrived this morning in time for breakfast, then off with Yvonne to Chris' place. I spent most of the day on the chapter for “Beautiful Architecture”, then out into the garden. It's warm again—today we had 35°.

Barbara brought some DVDs of people with problems training horses with her, which I was to copy. On a not completely related topic, spent some time looking once again at the issues surrounding DVD mastering and rediscovered any2dvd, which I had ported and then discarded a bit over a year ago. It now has documentation, so maybe I should resurrect it.

Also had fun with my PHP scripts for my diary. One of the instances of the showphoto function creates a simple image with a link to a larger version; the larger version includes a table to describe the photo. In addition, the small version displays photos side-by-side where possible, whereas the larger version always shows them arranged vertically. I had had the following code for some of the photos for last Saturday:

      <p>
      <?php
        showphoto ("Photos/20040205/Treetop-walk-9.jpeg");
        showphoto ("Photos/20071010/otways-7.jpeg");
        showphoto ("Photos/20080301/treetop-walk-14.jpeg");
       ?>
      </p>

Today for the first time I tried putting that page through the W3C validator and specifying the size; it failed dramatically with scope problems. It didn't like the <table> inside a <p> element. It didn't like a <br> either. It took me quite some time to discover a way to placate the syntax:

      <p/>
      <?php
        showphoto ("Photos/20040205/Treetop-walk-9.jpeg");
        showphoto ("Photos/20071010/otways-7.jpeg");
        showphoto ("Photos/20080301/treetop-walk-14.jpeg");
       ?>

That's truly a triumph of syntax over semantics.

Chris over in the evening, which we spent watching some of the DVDs.

One of the advantages of having visitors is that they normally want to go riding, which means that I finally get enough push to go out too, and so we did so today:

More work on PHP in the afternoon, finally cracking the issue of search paths for include files. Superficially that's straightforward enough: by convention they're defined in php.ini to be $_SERVER["DOCUMENT_ROOT"]/php/includes, an absolute path for the system. But like most web sites, www.lemis.com is a virtual host, in this case one of many on ozlabs.org, but there's only one php.ini file. Asked around, but nobody knew how to set the paths on a per-virtual host basis until Rasmus Lerdorf popped up. It's straightforward enough, once you know. In my local php.ini I have:

include_path = ".:/home/grog/public_html/php/includes:/usr/local/www/data/php/includes"

By contrast, on ozlabs.org I have:

<VirtualHost *:80>
        ServerName      www.lemis.com
  ...
        php_value include_path .:/home/grog/www.lemis.com/grog/php/includes
</VirtualHost>

The paths are different, of course; that's the main problem. But it's also important to notice the different syntax.

While looking at this, discovered a surprising number of messages in the error log of the form:

Wed Mar 05 09:20:24 2008] [error] [client 98.223.149.42] File does not exist:
  /home/grog/lemis.com/grog/Photos/20020220/Photos, referer:
  http://www.lemis.com/grog/Photos/20020220?S=A

Where does this ?S=A come from? I don't use anything like that on the web site. Looking at the access log, I found:

c-98-223-149-42.hsd1.in.comcast.net - - [05/Mar/2008:09:18:28 +1100] "GET /grog/Photos/20020220?S=A  HTTP/1.1" 301 440 "- " "Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 5.1; .NET CLR 1.1.4322;  InfoPath.2; .NET CLR 2.0.50727; .NET CLR 3.0.04 506.30; .NET CLR 3.0.04506.648)"
c-98-223-149-42.hsd1.in.comcast.net - - [05/Mar/2008:09:18:29 +1100] "GET /grog/Photos/20020220/Photos/20020220/small/sharpe-house.jpeg HTTP/1.1" 404 445 "http://www.lemis.com/grog/Photos/20020220/?S=A" "Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 5.1; .NET CLR 1.1.4322; InfoPath.2; .NET CLR 2.0.50727; .NET CLR 3.0.04506.30; .NET CLR 3.0.04506.648)"

I have no idea where the user got this first URL from, but /grog/Photos/20020220?S=A would refer to a file that's really there, in fact identical to the file in /grog/Photos-20020220.html—and with incorrect links, as the second log file entry shows. The correct URL was /grog/Photos/20020220/small/sharpe-house.jpeg (now moved to /grog/Photos/20020218/small/sharpe-house.jpeg) but the URL in the link was relative to /grog, not /grog/Photos/20020220, so it got duplicated.

I've been aware of the problem, but I didn't think that anybody knew these URLs; possibly somebody (Richard Sharpe?) had derived the URL from a photo URL. Set to fixing the problem, but it's not as easy as it seems: if I remove /grog/Photos/20020220/index.html, the server simply presents a directory listing. What I need is a new index.html file with a redirect to the correct place, which since the reorganization of my photos is no longer even /grog/Photos-20020220.html, but /grog/photos/Photos.php?size=small&dirdate=20020220. That proved to be simpler than I thought. First I created a generic redirect script, and then inserted customized versions with

=== grog@dereel (/dev/ttyp8) ~ 162 -> cd ~/public_html/Photos
=== grog@dereel (/dev/ttyp8) ~/public_html/Photos 163 -> for i in  [12]*; do
> rm -f $i/index.html
> sed < rawindex >$i/index.html "s:YYYYMMDD:$i:"
> done

My chapter for “Beautiful Architecture” is due in on 9 March, whenever that may be; I'll take that as being (US) Pacific Time, which makes it 10 March here. I now have all the material I want to present—I think—but now comes the tidy-up, and the way I write that's considerable. Spent a lot of the day doing that.

Also spent some time looking for the source of messages in my Apache 2.2 error log files (yes, they're in a directory reflecting the release of Apache):

[Wed Mar 05 00:16:52 2008] [error] [client 65.214.45.101] mod_mime_magic: invalid type 0 in mconvert().

After much searching, found that it appeared to occur when somebody accessed an EXIF file, which contains exposure data for JPEG photos. It would have been nice if the error message had referred to a file name. The problem is, of course, the nomen est omen issue: like Microsoft, Apache assumes that the name of a file, more specifically the ending, indicates the content, and it doesn't know what files with names ending in .exif are. A little more searching showed that there's an Apache directive AddType that specifies MIME types for files whose names aren't recognized, so I added this to the VirtualHost entry:

<VirtualHost *:80>
        ServerName      www.lemis.com
  ...
        php_value include_path .:/home/grog/www.lemis.com/grog/php/includes
        AddType         text/plain .exif
</VirtualHost>

And that stopped the problem. But it's tacky: text/plain should be the default for names that aren't recognized. Why isn't there a DefaultType directive? No good reason: in fact, there is a DefaultType directive. Added that to my config file, replacing the AddType entry, and tried again. The error messages came back! For some reason, Apache wasn't honouring the directive. More head-scratching; I'll look at that some other time.

Last week I wrote a letter to Prithie Naidoo of Telstra accepting their conditions and asking them to cancel the contract immediately and confirm. Of course, she didn't, and I was able to confirm that the account was still active. sigh Why are these people so terminally incompetent? Wrote another, more stiffly formulated letter giving Telstra until the end of the week, and rather pointedly stating that I was available on the phone, despite her claims to the contrary.

Yvonne is doing a mini-“clinic” next week, an Australian word for horse training, and we're feeding the group on one evening. After some discussion, we decided on chile con carne, whatever that means, and I'm the one to do the cooking.

Everybody knows that “chile con carne” means “beans with chili”—or do they? Of course, in Spanish it means “chile with meat”, and my research shows that in Texas, where the dish originated, that's exactly what they mean. No beans. There's a famous recipe by Lady Bird Johnson, US First Lady, Texan and wife of president Lyndon B Johnson.

There's a problem here: I'm looking for a dish with beans. LBJ's recipe is in fact quite boring in appearance: meat, onion, garlic, oregano, cumminseed, tomatoes, chili. If it weren't for the oregano, it would be a poor man's beef curry, not what I'm looking for. In the end I decided on a compromise between the LBJ recipe and the recipe in “The Joy of Cooking”, which does include beans. Out of deference to my Texan friends, who are many and burly, I'm calling it chile con frijoles. And to my surprise, the first time turned out pretty much the way I expected it to be, though I suspect that home-grown oregano is much weaker than commercial dried oregano, and that I should have used more.

Busy day. Started with problems measuring my blood sugar this morning: my sugar meter powered on with the message “Error 105”, and wouldn't reecover. Pulled out the batteries and replaced them, to no avail, and after establishing that there was no easily visible way of resetting the box, turned to Google for help—and found it almost immediately. Found the complete instruction manual on line, along with a searchable HTML version where I found:

Notice: In the unlikely event that your sensor displays an Error 105 message, simply remove the batteries, leave them out for 5 seconds, and reinstall the batteries. This will reset the sensor to function normally.

That worked. So I just hadn't left the batteries out long enough. But this gives rise to a number of questions:

• If this is so unlikely, why does the manual mention how to handle this error number, and no others (apparently there are several)?

• Why doesn't the manual say what the error numbers mean? I suppose this is the difference between users and people who understand the technology.

• On the face of it, this is something like a panic. I suppose that's a sign of the times, but can't the problem be handled better?

Had barely finished that when I got a call from Prithie Naidoo of Telstra, telling me that the service had been cancelled some time last week. I got her to check, and of course it was still active. I then asked her why she would not refund the modem, and she repeated the nonsense she wrote in her letter of 25 February. I suggested that she had not read my reply, and of course she claimed she had—but was not able to mention what I had said, let alone (obviously) come up with a counterargument. At one point she said “I know this has taken a lot of your time. It has taken a lot of mine too”. I had to point out to her that, unlike me, she was being paid for it. In the end, with help of the mention of formal complaints to the ACCC, and that the accounting is all wrong, she agreed to refund the modem, the fees from 1 February on (“please contact me if you have any trouble with the accounts”, presumably because she couldn't see any billing for that time), and the sums offered in her letter of 7 February. Hopefully that'll finally be an end to it.

Into town to do a number of things, including finally getting the alternator on the Commodore fixed, at John Stevens Auto Electrical in Redan. Also to Big W to complain about some of the bulbs I had bought last week, which were completely dried out—7 out of 30 bulbs. I suppose it's typical of companies like Woolworths that they replaced exactly 7 bulbs, by opening a pack of 16, rather than just give me the bag. No apology, no recognition of the fact that their failure had caused me problems, and I'm sure they'll just throw away the rest (of course, maybe the personnel will get them, which would at least make some sense).

Did a lot of other things, and by the time we had finished shopping, the car was ready—that was nice of John, and saved me a separate journey into Ballarat to pick up the car.

Back home and went to get something out of the car. Pulled the keys out of my pocket and—they weren't mine! Somehow I had accidentally taken somebody else's keys when picking up the car. And so I had to turn around and bring them back. What a waste of time!

Finally back home, found that my Ethernet switch had hung itself up—that happens far too often—and needs a power cycle. I rely too much on network connectivity for that to be acceptable. This time I lost my beer fridge, the MythTV backend on ceeveear and lots of windows from dereel displayed on eureka. Got that back together and started copying MPEGs from ceeveear to teevee—and the connection hung up. That's a separate bug in Linux (running on ceeveear): it seems to be far too common for Ethernet interfaces to wedge. I have a workaround:

while :; do
  ping -c 1 -W 1 192.109.197.137
  if [ $? -ne 0 ]; then
    ifconfig eth1 down up
    route add default gw eureka
    logger Bloody eth1
  fi
  sleep 1
done

But the switch hang had stopped the process, so the interface hung and required a reboot. What a pain!

Another bloody power failure woke me early this morning; this time it was different, though, because Yana and Sundance were due to arrive in Ballarat at 05:15, and Yvonne had set the alarm on her bed, which was now reset. Set my watch and gradually drifted off to sleep, only to find something crawling over my hand. Up like a shot, and turned on the light to find something dark in the shadows. Prodded it with something and it jumped away, so it wasn't a spider. Back to sleep, after a fashion. Yvonne later discovered it:

I wonder how it got there.

My switch wedging yesterday had further repercussions: my satellite statistics gatherer also stopped working. I really need to make these scripts more robust.

More work on the “Beautiful Architecture” chapter today, not much, but I'm now a lot happier with the text. I should make the deadline with no problems.

Indian dinner in the evening, Sundance and Yana helping. We eat “tortillas” as roti, but today Sundance and Chris found religious implications. Yana didn't, so Sundance cleaned out her brain:

Yana and Sundance set off today, slowly, to visit people they knew. At my suggestion, Yana called up my aunt Freda and cousin Gill, where they had intended to stay tomorrow night, so that they could be aware of their luck, but Gill wasn't available.

Printed out some Google Maps of potential itineraries, learning much in the process:

• Both firefox and Apple's safari print web pages in portrait mode and offer no obvious way of changing the orientation, even when the window is wider than it is high, which is normal for full-screen displays.

• firefox can't print out the display page correctly: the route (normally in blue) is missing. Since the step-by-step itinerary also doesn't indicate the towns, this results in a printout that is more guesswork than anything else.

• safari does print out the route, but it makes arbitrary and apparently non-overridable decisions about the mapping of the screen to the page, in this case chopping off the right half of the display. My guess is that something is assuming a maximum window size of 1024x768.

• Google does offer a “print” option on the page. Choosing that loses the map (unless you select it again, in which case you get a small version) and shows you detailed maps of the first and last step only.

• I thought that the breakage round Dereel was bad, but the directions from Healesville to Briagolong were fascinating: apart from stopping 200 metres before Briagolong, it changed completely if I specified the exact address. Instead of taking me to 556 Dargo Road, Briagolong, it took me to the Wellington Shire Council, 73-81 Macalister St in Sale, 32 km away on the wrong side of Briagolong. This one completely baffles me.

Yana and Sundance finally got on their way in mid-afternoon, headed for Healesville, leaving me time to create an enormous pot of chile con frijoles, which in recognition of Pedernales River Chili, I've decided to call Dereel chile. I've also decided that our cooker is just not hot enough to handle this stuff; the juice runs out of the meat and boils off.

And I thought summer was over! Today the temperature hit 40°, and they're promising a continuation of the weather. It's strange, though, where the horses are changing coat and the deciduous trees are going yellow.

Phone call from my cousin Gill this morning: she hadn't been able to get back to Yana, but she was expecting many friends to come in from the USA, and it wasn't convenient. This seems to be the month for overseas visitors, but it also should maybe bring home to Yana that planning is a good idea.

Today or tomorrow is the deadline for the “Beautiful Architecture” chapter, and spent most of the day working on it. It's sort of finished now, though a second lot of proof reading, with others helping, would certainly help.

PHP is not C, part 323

Spent a lot of time trying to do some simple string processing in PHP today. PHP has a feature called a heredoc, which looks surprisingly like the Bourne Shell's here document (note the different spelling). In the Bourne shell, you can include a document directly in the shell:

=== grog@eureka (/dev/ttypa) ~ 3 -> ./myprog << EOF
foo
bar
baz
zot
EOF

From within the C program, you can then use gets() or fgets() to read individual lines.

PHP has a deceptively similar construct. For example, you can write:

print <<< EOF
foo
bar
baz
zot
EOF;

And yes, the <<< are correct; << is a shift operator, so it can't be used here. But how do you read it? There's no gets() function. There is an fgets() function, but for that you need to open a file. Where's the file here?

After further examination, discovered that a heredoc is quite a different animal from a here document: it returns a single string, \n characters and all. So I set to parsing the string and found it incredibly difficult. Asked Rasmus:

<groggy> Rasmus: OK, I give up.  How do I process a heredoc line by line?
<groggy> There must be a way, but I can't find it in the docco.
<Rasmus> I don't understand the question.  Do you just want to split a long
     heredoc string on \n ?
* groggy is Enlightened.
<groggy> OK, rephrasing helps a lot.  Thanks.
<Rasmus> see php.net/explode
<groggy> Rasmus: Yup, just what I'm doing.
<groggy> Rasmus: Just the C mentality showing through :-)

It's easy when you know how: explode ("\n", $mystring) returns an array of the lines in the heredoc that you can then process with foreach. But C it's not.

Chris Yeardley along for dinner in the evening. After all the work I went to to ensure that selecting a photo on a web page (this one, for example) gives a bigger version in context, it proves that she preferred the old method, where it went straight to the day's unformatted photos. You can't please everybody, but I thought that I might add a “legacy” mode to the pages

Mail from Ian Donaldson this morning talking about how to print in landscape from an Apple. After some comparison, found that his description applied only to Mac OS 10.5, which doesn't even run on my box. On 10.4 there still doesn't seem to be any way to do it. Even under 10.5, it seems that there's one default, regardless of the orientation of the original. Sent Ian back some screen shots, in the process first finding out the magic key combination Apple-Shift-4, after which you get a cursor that you can either use to define a section of screen or position on a window and press space, then the inevitable mouse click. It creates an image with a name like ~/Desktop/Picture 1.png, with these horrible embedded spaces to make them more difficult to process with scripts, and placed not in the Pictures directory, but in the Desktop directory. Finding files on “modern” computers seems to be based on luck and much searching. Whatever happened to the good old path name?

More frobbing my “Beautiful Architecture” chapter; a couple of people have reviewed it for me, and that seems to have improved it. Nothing from my editor, though. The sources are being stored in a subversion repository at O'Reilly, so I can see what the other authors have done. Most haven't committed anything yet, though today's the deadline. It looks as if I'll have plenty of time for correction and conversion to DocBook.

I use MythTV to record TV programs, but mplayer to view them: I haven't been able to get MythTV to handle XvMC correctly, and the user interface is also pretty terrible, while I can mainly frob mplayer to do what I want. But one thing that mplayer doesn't have is commercial skip. Today spent a bit of time investigating how MythTV does it:

• There's a program called mythcommflag that reads the entire MPEG file and uses various techniques to recognize the commercials. It doesn't have a man page (of course), but it does have a --help option that prints out some information.

• It's a standalone program, but it seems to require a database entry for the file:

  === grog@ceeveear (/dev/pts/3) /spool/Oldimages 76 -> mythcommflag -f 1007_20070504115500.mpg
  2008-03-12 10:51:04.466 Using runtime prefix = /usr
  2008-03-12 10:51:04.567 New DB connection, total: 1
  mythcommflag: ERROR: Unable to find DB info for 1007_20070504115500.mpg
  

• mythcommflag can generate a “cutlist”, which looks like this:

  === grog@ceeveear (/dev/pts/3) ~ 79 -> mythcommflag -f 1005_20080309210000.mpg --gencutlist
  2008-03-12 10:52:26.891 Using runtime prefix = /usr
  2008-03-12 10:52:26.950 New DB connection, total: 1
  2008-03-12 10:52:26.973 New DB connection, total: 2
  === grog@ceeveear (/dev/pts/3) ~ 80 -> mythcommflag -f 1005_20080309210000.mpg --getcutlist
  2008-03-12 10:52:30.842 Using runtime prefix = /usr
  2008-03-12 10:52:30.895 New DB connection, total: 1
  2008-03-12 10:52:30.921 New DB connection, total: 2
  Cutlist: 13358-15003,32856-38124,54805-63261,71539-77200,92037-98236,133326-141841
  

What are the numbers? Not documented, of course. Looking at the values, decided that they could be 32 kB DVD data blocks. That doesn't have much to do with MPEG-2 per se, of course, but why should that stop it happening? Spent a fair amount of time writing a copy program that would read in a cut list and skip it, made no easier by the fact that this was for Linux, and Linux still requires special treatment for files larger than 2 GB. I recalled that, but forgot the details. OK, that's what man pages are for:

=== grog@ceeveear (/dev/pts/3) ~ 1 -> man 2 open
No manual entry for open in section 2
See 'man 7 undocumented' for help when manual pages are not available.

35 years ago, one of the great advantages of the UNIX system was that it came with online documentation. Why is it that nowadays, where disks are up to a million times as big as on the early UNIX systems, it's still an optional extra? And where do I find it? With a bit of help from Jeremy Kerr, discovered that for Debian it's a package called manpages-devel. Installed that and got things working. The operative difference is the flag O_LARGEFILE to open, but by default it's not defined. First you have to tell your program that you want to use it:

#ifdef linux
/* First part of jumping through hoops to get open to work correctly */
#define _LARGEFILE64_SOURCE
#endif

In addition, the function names have changed, so all over the place in the program, you have things like:

#ifdef linux
  if (fstat64 (infile, &statbuf) < 0)
#else
  if (fstat (infile, &statbuf) < 0)
#endif

All this, it seems, is to maintain compatibility between older library versions and newer kernels. Why not just require that they match?

After doing all that, found that the cuts were in the wrong place. Compared with mythfrontend and discovered that, for the file I was looking at, at any rate, the “blocks” were about 38,800 bytes long—clearly not real blocks. After a bit of further comparison of the numbers, discovered that the numbers aren't blocks at all: they're frames. And we have programs for that, at least transcode. So pretty much a waste of time. Now I just need to read and understand the transcode man page (yes, it has one, 3174 lines long).

Instead, turned my attention to my budget. I still haven't found any good software to give me an overview of my expenditure, and for some time I have been mulling over how to do it in PHP. Today was the day, and by evening I had something in place, admittedly ugly, but I've broken the back of the problem, and the rest will be detail work.

More correspondence with Ian Donaldson: he actually went to the trouble to install MacOS 10.4 on a machine to find how to specify display orientation. It's in the page setup menu, accessible from the Safari File menu or with Shift-Apple-P. But in my situation, you won't see that: there's a “print” link on the page itself, and that just takes you to the print menu. And the default shouldn't override the natural orientation of the image, as it seems to do even in the most recent MacOS.

Also mail from Michael, without surname, telling me how to configure the screen saver to save in a more logical directory:

Open a terminal window as the user in question (not root)

type: defaults write com.apple.screencapture location /Users/[username]/Pictures

Then Type: killall -HUP SystemUIServer

or just log out and back in again.

The changed default is now stored in ~/Library/Preferences/com.apple.screencapture.plist

I'm sure there is a way of changing the name, they type and the quality, but haven't had to look for that before.

That's more than I wanted to know for the screen saver, but the presence of the defaults program (and a man page titled “BSD General Commands Manual”, which shows it to be part of the defaults system) is of great interest.

Somehow the main topic of the day has been documentation. Where's the man page for mythcommflag? Why no man pages on Linux? How do I find out how to do things on Apple? The second issue, of course, is how Apple sets its defaults. Why does it write screen captures to the Desktop directory? Why does it put spaces in the names? They're not necessary, and they make it difficult for Apple's tools to handle things.

More work in the garden today; the weather has been unseasonally hot for too long, and it's showing no sign of letting up. While looking around, found yet another surprise:

No idea what it is, but it was growing where I hadn't even expected flowers and where I hadn't watered. Clearly a survivor.

Peter Jeremy later sent me mail saying that this is a Gazania

While weeding another bed, managed to break off part of a similar flower. I had been planning to try cuttings of all this kind of flower once the weather got cooler, but it was there now, so decided to plant a few anyway and see how they turned out:

From left to right, the front row should become the following plants:

Interestingly, all of these plants survived except for the Cape Daisies (Osteospermum ecklonis), which are pretty indestructible as long as they get enough water. In particular, the orange Gazanias did very well, flowering profusely from October 2008.

Today was the first day of a “clinic” that Yvonne and Chris were holding from some people from the North-East of Victoria, and the inaugural Dereel Chile. We came well prepared, and it was a complete success:

Why is it still so hot and dry? Last month was cooler, and we had over 100 mm of rain. So far this month it's been 1.5 mm.

Nonetheless, spent some time tidying up the shed, with temperatures in the low 30s. Some day soon our dining room might be usable again.

David Gibson was having trouble with an online passport application: it came with a PDF that wouldn't print correctly. Mine expires in less than 2 months, so decided to try it out, despite a horribly slow connection to the passport site. To even get a renewal form I first needed to fill out a form, in three pages: first, only the passport number; then personal details (the sort of thing that you can find on my home page and elsewhere on this web site); and finally, my mother's maiden name. That's the only piece of information not available to somebody who has seen my passport title page. So I entered all this, taking about 2 minutes for each page change. “The data does not match a document”. Was it maybe the date of birth, which they wanted entered as a digit, in contravention of the Australian Government (Ausinfo) style guide, and which I entered without leading zeros? No, still didn't work (assuming the presumably broken program noticed the re-entry). Third time I was locked out, presumably meaning that the data was still not accepted. Called up the phone number, spoke to Tim, who confirmed that the information was correct, but when I asked him to take a complaint about the web site, he hung up on me. What are our authorities coming to?

One thing I couldn't blame on the passports web site was the speed of the transfers: it proved that my satellite downlink was seizing up again. TCP stopped working almost completely. It seems that it's yet another problem with this horrible “NettGain” BST implementation. Maybe they call it “NettGain” because so many negatives detract from the advantages.

I've mentioned the SBS Food Safari on a number of occasions in the past. Nearly a year ago, I accidentally deleted the “Malaysian Safari”, which particularly annoyed me, since it contained a recipe for nasi lemak. They're repeating the programs now, and today I finally watched the program. I'm not really much happier than when I deleted it. These programs could be good, but they're so inaccurate and misleading that they're hardly worth the trouble. At the time, the choice of recipes was acceptable if not typical. Since then, they've presented programmes on Indonesian and Singapore cuisine, both very similar. But all but one of the recipes could better be called Indonesian or Singaporean:

• First came the nasi lemak recipe, which was really just the rice. He uses much more daun pandan than I do, and also adds ginger, which I haven't seen elsewhere. Then he cooks it in a rice cooker. In my recipe I write: “cook until the liquid has been absorbed, stirring all the while to avoid the coconut milk burning on the bottom of the saucepan”. You can't do that with a rice cooker, and indeed the rice was burnt when he took it out of the cooker.

  But is this a Malaysian dish? My Singaporean friends would say that it comes from Singapore, and that's my impression, though most of the new pages claim it's Malaysian. Certainly it's easier to find in Singapore. And Food Safari had a separate programme on Singaporean food.

• Then there was the rendang, notable because the cook took red onions and called them “Spanish onions”—and that's what you'll find in the online recipe. This one looked good, but then completely disqualified itself by adding pre-packaged “curry powder”, whatever that may be. In any case, as I've been told in no uncertain terms by a character called Caniago,

  In Malaysia rendang recipes are often (usually?) corrupted by Indian and Chinese influences, so anything is possible over there. "Authentic" rendang, the type you will find cooked in Minangkabau villages in Sumatra...

  I disagree with him on the first statement, but it's clear that rendang is Minangkabau and thus Indonesian.

• Next was a roti canai, which I suppose you could call Malaysian, though I can't recall ever having seen it when I was a kid—instead we saw the various Indian rotis from which it descended, such as paratha or murtabak. The cook greatly irritated me by calling it just roti, which is the Malay word for bread, as the link confirms, and by asking for olive oil. Olive oil? What's that? Certainly nothing Malaysian, and with a strong enough flavour to be completely wrong. The recipe on the web site corrects that and suggests peanut oil, which is quite common in Malaysia. But it still looks just plain All Wrong. No egg. No fat. By contrast, the Wikipedia page states:

  The dish is composed of dough containing copious amounts of fat, egg, flour and water. The form of fat used is usually ghee (clarified butter).

  I'm surprised that the cook was prepared to have his name and the name of his restaurant at the top of the recipe. Still, roti canai has evolved enough to have some justification in calling itself a Malaysian dish. I just can't imagine anybody making it at home.

• Next was Char Kway Teow. It comes from Singapore, though I suppose it's pretty popular in Malaysia nowadays.

• Finally, there was a laksa recipe. As I mention in my recipe, there are two versions: Penang Laksa and Singapore Laksa. The cook for this version comes from Penang, so you'd expect Penang Laksa, wouldn't you? And be disappointed. His recipe uses coconut milk, a clear giveaway for Singapore Laksa. Penang Laksa is dominated by the sourness of the tamarind and the aroma of the mint and raw onions.

The whole programme doesn't even mention that the richness of Malaysian food comes from the fact that it includes three disparate cuisines: Indian, Malay (including Indonesian) and Chinese. Some of the statements in the ingredients list are strange:

• Thick Caramel Sauce - even more viscous than dark soy, this is used to add a depth of colour and flavour to many dishes.

  I've never heard of this before. It really seems to be just caramel, and the bottle shown in the film shows that it was made for Campbells, a Western company. I suspect that this is comparable to adding tomato ketchup to everything.

• Pandan Leaf — is the Asian equivalent of vanilla essence.

  I don't know what to say to this one, apart from the fact that they seem to have confused “Asian” and “Malaysian”.

In summary, I find more and more that the Food Safari could be quite good with just a little more research. I've found the programmes on cuisine that I don't know to be reasonably good; clearly I'm being misled. And wouldn't it be nice to get a list of shops who sell raw materials for non-English foods in Australia?

A letter in the mail from Telstra today, enclosing a cheque for $901.93, without further comment. The cheque was dated 10 March (Monday), but the post stamp was Wednesday, 12 March, the day I wrote my last letter asking Prithie to clarify things. I didn't get a clarification, and in principle the cheque would be good enough—except that it's not for the full amount: a sum of $8.24 seems to be missing, not to mention two months' fees. Why are these people so difficult to deal with?

The weather has been stinking hot for over a week, and it's showing no signs of letting up. Today it wasn't just hot—in the high 30s—but there was a very strong, unabating wind from north and west which made it all the more unpleasant and dried out the plants, not to mention the bushfire danger. And that in the middle of March! Hopefully it'll cool down soon.

Spent a lot of time finalizing details of a bassoon swap with Aligi Voltan. He wants to send them with UPS. Hopefully they'll make it here without problems; based on my prior experience, I'll be lucky. Spoke to Cathy at UPS, who told me that I would not be able to clear customs myself, since they offer a “door-to-door” service, and where that's not possible, it's sent to the nearest post office by Australian Air Express. I wonder if she believes that.

Into the office today to discover that Aligi has already sent off the bassoons—via bloody UPS, despite my warnings. He states, quite probably correctly, that you can't trust the Italian postal service. Spent some time tracking them—the display looks good, but I've already observed that it's not completely reliable. Still, I suppose I can believe it when they tell me that by the afternoon they had left Köln for an unspecified destination. I wonder if they'll go the long way round the world.

Spent some time registering for UPS web site, which was the usual mess. After about 20 minutes registration I discovered that they had nothing useful to say.

The hot weather continues. The combination of that and the high winds yesterday has completely dried out the bird bath. Here “before” and ”after” images showing one day's evaporation:

Spent some more time investigating MythTV cut lists today, by the expedient of building one and seeing which tables in the database changed their access time. It turned out to be recordedmarkup, and for some reason the key isn't the name of the file, but the “channel” number and start time:

mysql> select * from recordedmarkup order by starttime, mark;
+--------+---------------------+--------+--------+------+
| chanid | starttime           | mark   | offset | type |
+--------+---------------------+--------+--------+------+
|   1005 | 2008-03-12 14:00:00 |      0 | NULL   |    4 |
|   1005 | 2008-03-12 14:00:00 |      0 | NULL   |    1 |
|   1005 | 2008-03-12 14:00:00 |   1831 | NULL   |    5 |
|   1005 | 2008-03-12 14:00:00 |   1831 | NULL   |    0 |
|   1005 | 2008-03-12 14:00:00 |  35288 | NULL   |    4 |
|   1005 | 2008-03-12 14:00:00 |  35288 | NULL   |    1 |
|   1005 | 2008-03-12 14:00:00 |  40058 | NULL   |    5 |
|   1005 | 2008-03-12 14:00:00 |  40058 | NULL   |    0 |
|   1005 | 2008-03-12 14:00:00 |  55491 | NULL   |    1 |
|   1005 | 2008-03-12 14:00:00 |  55491 | NULL   |    4 |
|   1005 | 2008-03-12 14:00:00 |  57375 | NULL   |    0 |
|   1005 | 2008-03-12 14:00:00 |  57375 | NULL   |    5 |
|   1005 | 2008-03-12 14:00:00 |  91307 | NULL   |    4 |
|   1005 | 2008-03-12 14:00:00 |  91307 | NULL   |    1 |
|   1005 | 2008-03-12 14:00:00 |  92822 | NULL   |    5 |
|   1005 | 2008-03-12 14:00:00 |  92822 | NULL   |    0 |

mark is clearly the frame number, and it looks like type 4 is start of cut and type 5 is end of cut. There are also -3, 0 and 1, which I haven't deciphered yet.

Spent some time trying to find an appropriate transcoder. Why are all these programs so difficult to use? In the end, settled on mencoder—the online documentation is getting reasonably good. Created a little script to extract an Edit Decision List (EDL) from the database and use it to re-encode a programme, but I must have missed a parameter: the time information was wrong, and so the skips were in the wrong place. That doesn't seem to be related to the EDL: I had the same problem without. More investigation needed.

I'm thinking of buying some new lenses for my camera, in particular a 650-1300 mm extreme telephoto zoom, so decided to fake some photos with those focal lengths:

Image title: horses 14mm          Dimensions:          3684 x 2736, 5181 kB Make a single page with this image Hide this image Make this image a thumbnail Make thumbnails of all images on this page Make this image small again Display small version of all images on this page All images taken on Sunday, 16 March 2008, thumbnails          All images taken on Sunday, 16 March 2008, small Diary entry for Sunday, 16 March 2008 Complete exposure details

 

Image title: horses 42mm          Dimensions:          3648 x 2736, 4513 kB Make a single page with this image Hide this image Make this image a thumbnail Make thumbnails of all images on this page Make this image small again Display small version of all images on this page All images taken on Sunday, 16 March 2008, thumbnails          All images taken on Sunday, 16 March 2008, small Diary entry for Sunday, 16 March 2008 Complete exposure details

   

Image title: horses 650mm Complete exposure details Dimensions: 600 x 450, 117 kB Dimensions of original: 600 x 450, 117 kB Display this image: thumbnail    hidden   alone on page Display all images on this page as: thumbnails    this size Show for Sunday, 16 March 2008: thumbnails    small images    diary entry

 

Image title: horses 1300mm Complete exposure details Dimensions: 600 x 450, 85 kB Dimensions of original: 600 x 450, 85 kB Display this image: thumbnail    hidden   alone on page Display all images on this page as: thumbnails    this size Show for Sunday, 16 March 2008: thumbnails    small images    diary entry

The last ones aren't very good, of course, but that's the reason for the new lenses. But how to make the magnified photos? My tool of choice is xv, but it doesn't give me the choice of setting exact bounds, so tried the gimp once again.

How do you crop? It's pretty non-intuitive, so I pressed the Help button. No help! My first question was “what the *(&*& were these people thinking to install a port without documentation?”. Then I checked and found that the GIMP documentation comes in 13 languages in a 50 MB tarball, and it installs all of them!:

===>  Building for gimp-help-0.12
*** Creating directories ***
Creating base image dirrectories.
mkdir: html/images/.: File exists
Creating image dirrectories for language: cs
Creating image dirrectories for language: de
Creating image dirrectories for language: en
...
Creating image dirrectories for language: zh_CN
...
===>   Registering installation for gimp-help-0.12
     5470.67 real      2036.42 user        72.84 sys
     95260  maximum resident set size

And yes, the misspelling is original. There's plenty of room for improvement there.

Following my bassoon transport showed that they did, indeed, come east, via Dubai and Bangkok to Melbourne, where they arrived at 14:39 and immediately went into exception state. I think that means “sent to customs”.

Took some photos of the last time I'll play my 1847 Savary, but while reading them into the machine, it crashed. When it came up, it had the worst file system damage I have ever seen; most of /etc was gone, what was left contained severe directory damage, and the disk label on /home had been overwritten. Discovered that the man page for fsck no longer tells you about the alternate superblocks—that's now in fsck_ffs, and as I correctly recalled, it's -B 32 for UFS, and -B 160 for UFS2.

While I was tidying up there, Arnd Bergmann showed up. Left the disk restore run overnight.

Up this morning to continue the recovery of dereel's file systems, which on and off took most of the day. It's clear in retrospect that my dual root file system approach to system upgrades is also of immense value in recovering trashed file systems. So was my mklinks program, which I should finally make into a port.

But what happened? Files disappeared at random. The photos I created yesterday with the gimp survived, but some of the originals disappeared. Looking at the output of fsck as it shot by, it seems that whole groups of inodes were trashed. So, it seems, was the disk label on /home. Callum Gibson suggested that I might have had less damage had I stuck to the traditional file system layout (separate /, /usr and /var), and mounting the root file system read-only, but I don't think so: the nature of the damage suggests random scraping of the disk, probably overwriting groups of inodes, and that wouldn't have respected a read-only mount. Also, by definition, the read-only part is easy to recover: it'll all be in yesterday's backup.

I have good reasons for the current layout; I don't think that anything that happened here relates to it.

My bassoons arrived in Melbourne yesterday, and this morning, once I got my web browser working again, I checked what had happened, which appears to have been in reverse order:

  17/03/2008  9:22  A QUARANTINE WAS REQUESTED BY A GOVERNMENT REGULATORY AGENCY / RELEASED BY
                    CLEARING AGENCY.  NOW IN-TRANSIT FOR DELIVERY
  17/03/2008  7:56  PACKAGE DATA PROCESSED BY BROKERAGE.  WAITING FOR CLEARANCE / BROKERAGE RELEASED
                    SHIPMENT.  SHIPMENT IS SUBMITTED TO CLEARING AGENCY FOR FURTHER CLEARANCE
  17/03/2008  7:56  A QUARANTINE WAS REQUESTED BY A GOVERNMENT REGULATORY AGENCY
  17/03/2008  5:17  IMPORT SCAN

What happened at 9:22? Was it finally sent to AQIS, or did it get released then? Called up the UPS number (131 877) and spoke to Ben, a woman, who told me that the instruments had been released from quarantine at 9:47 this morning, which didn't match any of the times. She also told me that they would be delivered to me on Thursday, to the address on the docket. I asked her to confirm that, and that I had been told only on Friday that they would be sent to the nearest Post Office by Australian Air Express. She checked the post code and confirmed it. I'll still believe it when I see it.

Later Arnd came in with some of his photo gear, of which he has much, including an adaptor card to put SD flash cards in a Compact Flash socket. Tried that in my camera and managed to bend a couple of pins in the connector. Spent a worried half hour with a watchmaker's screwdriver trying to bend them straight again, and was finally successful, though they look a little wobbly. I think I'll leave the CF card in there permanently now.

Last night I had dragged out my old wireless access point, which I had discovered to be defective a couple of months ago. That interested Arnd, who established that it was a clone of the Asus WL-500G and decided to re-flash it with software from OpenWrt, with a surprising amount of success, including getting the wireless component to work again. Instead of a broken web interface I now have telnet access, and there's even a vi editor on the machine. I must take a look at that in more detail.

Chris came over for dinner—kangaroo steak on the barbie—and we ended up with our usual nonsense:

Arnd off at the crack of dawn today: his plane was scheduled to leave at 8:55, but it appears to have been delayed. On IRC I saw:

*** arnd (~konversa@203.94.133.193) has quit: Remote host closed the connection
                                    [09:05]

One of the things that interested me about his visit was that he came to terms very well with the software he uses, mainly KDE. He recommended digikam for processing photos: I've found limitations both with xv (loses exif information, and there's no way to specify a crop size in pixels) and gimp (slow, cumbersome and, at least currently, confusing), so yesterday I started to install digikam. Today, finally, the installation completed. The time wasn't just because of the 15 dependencies, including apparently half of KDE and also sqlite, but because many of these had configuration dialogues that held things up until I found them. Finally started it, getting lots of informational messages in the terminal window. I suppose that's better than GNOME's habit of presenting lots of (hopefully benign) error messages, but it makes me wonder if people know the use of the & character in the shell. But then, shells are out, mouse-pushing is in.

The first thing I got was the question:

Image title: digikam 1 Dimensions: 482 x 361, 41 kB Dimensions of original: 482 x 361, 41 kB Display this image: thumbnail    hidden   alone on page Display all images on this page as: thumbnails    this size Show for Tuesday, 18 March 2008: thumbnails    small images    diary entry

A folder? I don't want to fold my photos!

More seriously, though, this is the kind of problem that prevents me from using this kind of program—not the bad language, but the fact that the program is telling me how to organize my photos. After years of experimentation, I've come to store my photos in a directory relating to the date; for example, today's photos are in ~/Photos/20080318, and there I keep both the original and a reduced copy. On 10 March, for example, I only took one photo. The directory contains the original (P3102066.JPG), the processed copy in original size for the web site (birdbath-full.jpeg) and a couple of housekeeping files:

=== grog@dereel (/dev/ttyp3) ~/Photos/20080310 176 -> l
total 8
-rwxrwxrwx  1 grog  502   7107548 Mar 10 10:44 P3102066.JPG
-rw-r--r--  2 grog  home  1220929 Mar 10 10:44 birdbath-full.jpeg
-rw-r--r--  2 grog  home       23 Mar 11 09:27 description
-rw-r--r--  1 grog  home       85 Mar 11 09:27 script.45414.sh

For the web, I store the photos in three different sizes, along with more auxiliary files, most of which should go away when I get things stored in a data base:

=== grog@dereel (/dev/ttyp3) ~/public_html/Photos/20080310 182 -> l -R
total 1
drwxr-xr-x  2 grog  lemis   512 Mar 11 09:27 big
-rw-r--r--  1 grog  lemis  4185 Mar 11 09:28 birdbath-full.exif
-rw-r--r--  2 grog  home     23 Mar 11 09:27 description
-rw-r--r--  1 grog  lemis    56 Mar 11 09:28 photolist.php
drwxr-xr-x  2 grog  lemis   512 Mar 11 09:27 small
drwxr-xr-x  2 grog  lemis   512 Mar 11 09:28 tiny

./big:
total 2
-rw-r--r--  2 grog  home  1220929 Mar 10 10:44 birdbath-full.jpeg

./small:
total 1
-rw-r--r--  1 grog  lemis  81869 Mar 11 09:27 birdbath-full.jpeg

./tiny:
total 1
-rw-r--r--  1 grog  lemis  41620 Mar 11 09:28 birdbath-full.jpeg

One thing that is very important is space. The file birdbath-full.jpeg is 1.2 MB in size. I have it in two places; how do I avoid wasting this much space? The answer's in the directory listing, of course: look at the link count. The file ~/public_html/Photos/20080310/big/birdbath-full.jpeg isn't a copy of ~/Photos/20080310/birdbath-full.jpeg; it's identically the same file, as a display of the inode number shows:

=== grog@dereel (/dev/ttyp3) ~ 191 -> ls -li Photos/20080310/birdbath-full.jpeg \
                                      public_html/Photos/20080310/big/birdbath-full.jpeg
348 -rw-r--r--  2 grog  home  1220929 Mar 10 10:44 Photos/20080310/birdbath-full.jpeg
348 -rw-r--r--  2 grog  home  1220929 Mar 10 10:44 public_html/Photos/20080310/big/birdbath-full.jpeg

The inode number (first column) is the real file identifier. That works because the names, and not the contents, are stored in the directory. That's why “folder” is such a silly word. I don't use KDE, so I can't see what information it displays, but Apple's “Finder” can't find that information:

Image title: finder 2 Complete exposure details Dimensions: 265 x 826, 109 kB Dimensions of original: 265 x 826, 109 kB Display this image: thumbnail    hidden   alone on page Display all images on this page as: thumbnails    this size Show for Wednesday, 19 March 2008: thumbnails    small images    diary entry

I haven't had time to look at digikam in more detail (see below), and I have no idea if its view of the world even allows the kind of structure, but that's the point: I want a tool, not something that makes me change the way I work.

Somehow the whole day was taken up with bassoons. To start with, I needed to pack and send the 1847 Savary, including finding a transporter. Checking the Yellow Pages showed that the only other services in the area apart from UPS and Australia Post were DHL and Australian Air Express.

Tried them all:

• I had already had a verbal quote from UPS on Friday, but they have different tariffs, so went through things in (excruciating) detail on the web. Got back a detailed quote telling me that they couldn't pick up the parcel today (the time the web site had suggested), and if they picked it up tomorrow, they could deliver in on Thursday, 27 March (8 days), at 9:00 am for $547.09, at 10:30 am for $453.09, by the end of the day for $435.70. Alternatively, for only $404.68, they could deliver on Tuesday, 1 April.

• DHL is another company with an illegible web site, so I called up and was dealt with pretty efficiently by Naomi, who gave me a delivery time of 4 business days (end of 26th, presumably) for $390.31. The quote number she gave me was surprisingly short: 20. I wonder how much stuff they move.

• Kitty at Australian Air Express was new, so she put me through to Catherine, who put me through to the International Department, where I got an answering machine.

• Australia Post has a confusing web site, and I wasn't sure which service to use. Called up on the phone and navigated a correspondingly confusing phone menu system. You need to know what service you want to be able to use it. Chose one at random and was connected to Cameron, who was pretty on the ball and said that only Express Courier International (which, for some reason, is abbreviated to EMS) had tracking. That would cost a total of $166.70 and be delivered in Italy by a courier service, since the Italian Post Office is apparently not overly reliable.

Decided for the Australia Post solution, at least because they seemed to be the most knowledgeable. I could well imagine that UPS would have wasted another day when they discovered that they don't pick up in my area after all.

Next was packing the instrument, which took two tries. In each case we cut a cardboard moving carton to shape, but the first time didn't look too good:

I had agreed to send the instrument today, but if I had known that the others would be here so soon I would have been tempted to put it off until tomorrow so that I could have a photo of all 11 bassoons that I had in the house today, instead of only 10:

Into the office this morning to discover that my machines weren't responding properly, again. Worried that dereel had had another crash like the one on Sunday, but it turned out to be my main switch, which has given me trouble in the past. Today I was able to get it to work again, but only for 30 minutes, so it looks like I'm in for a new switch soon. Found an old one with a 115 V power supply and put it in temporarily.

Round breakfast, got a call back from Australian Air Express. Yes, they could transport my bassoon, price in the low $200 range. That's better than UPS or DHL, but given the timeliness I'm glad I didn't take them. To their defence, it seems that they deal more in freight than parcels.

One of the things that Aligi sent with the bassoons was a DVD+R with some of his music in normal directories and WAV format. That sounded like a job for the Apple, so set to to play them back.

But how? Presumably if you put a real audio CD in the player, it will be detected and you get an icon on the screen. But this wasn't a real audio CD. It was detected and mounted, of course, but the only information I got was a “finder” window. Clicking on the icons seems to do nothing unless you do it at least twice—even then it's difficult to tell because you get no visual feedback that it is off doing something. Went looking for programs (sorry, “Applications”) to play the data and found some that might fit the bill. But which? iTunes.app? QuickTime&nbsp;Player.app? Tried the latter and couldn't even work out how to start it. Clicking multiple times on the name changed the highlighting, but didn't start the application. Or maybe it did; at some random later time a window appeared titled “Apple QuickTime”. To judge by the window, it's a advertisement downloader. Clicked on the “About” tab on the top menu, but of course that's more bad language and really just tells you the release version. Would it be too much to ask to put a two-line description in the “About” box which states what the program is intended to do?

The Help isn't much better. It doesn't tell you what it does either; the first heading is “What's new?”, and it obviously expects you to know what it's good for. You can also “browse” the help, which involves selecting a potential topic. “QuickTime Player” sounded good, so I clicked on it. Submenu “What is QuickTime?” tells me:

QuickTime Player is a free multimedia player. You can use it to view many kinds of files, including ... audio...

Never mind what they mean by viewing an audio file; that seems to be what I'm looking for. But then it leaves you to find your way back. Reading through this “help” appears to require more mouse movement than eyeball movement. What are these people thinking?

Still, that's clearly the one to use. I wonder what iTunes.app is, but finding out is too difficult. With a bit of effort I can select the files I want to play—one at a time! What a pain.

Decided to burn the data to CDs instead. But how? All mastering seems to be such a pain. After a bit of investigation, came to the surprising discovery that, for audio CDs, no mastering is necessary if you have WAV files. Just give the names in sequence to your favourite burner program, and it'll create a CD.

That's the theory, anyway. First, of course, the file names were full of spaces, which the programs would interpret as file name delimiters. Wrote the following little bash function, which didn't work:

# Replace spaces in names with -.
# This doesn't work for the same reason that spaces in names are such
# a pain in the first place.
spaceout ()
{
  (for i in $*; do echo mv \"$i\" `echo $i | sed "s:':-:g; s:^:':; s: :-:g; s:$:':"`; done) | sh
}

Next, which burn program? FreeBSD comes with a program called burncd:

=== root@teevee (/dev/ttyp4) ~ 67 -> burncd -f /dev/acd0 audio `ls -1rt *.wav` fixate
next writeable LBA 0
writing from file Mozart---1-Kyrie.wav size 71490 KB
written this track 735 KB (1%) total 735 KB
Input/output error

next writeable LBA 472
writing from file Mozart---2-Gloria.wav size 22298 KB
written this track 735 KB (3%) total 1470 KB
Input/output error

What went wrong there? Who knows? Tried again with cdrecord which worked fine:

=== root@teevee (/dev/ttyp4) ~ 93 -> cdrecord dev=1,0,0 -speed=2 -pad -audio `ls -1rt *.wav`
...
Starting to write CD/DVD at speed 4 in real TAO mode for single session.
Last chance to quit, starting real write    0 seconds.  Operation starts.
Track 01: Total bytes read/written: 106919568/106919568 (45459 sectors).
Track 02: Total bytes read/written: 57492288/57492288 (24444 sectors).

The ls -1rt is there to order the files by their modification timestamps and not their names, which could be random, and the dev=1,0,0 is a strangeness of cdrecord, which wants SCSI addressing and not device names.

There are further strangenesses: I couldn't find any information about the speed of the CD-R blanks, so I chose 2 to be on the safe side; without any warning, it sets the speed to 4, almost certainly a good choice. It's also not very informative: after the message “Operation starts”, nothing happens. Looking at the device with iostat shows that it's inactive. In fact, that appears to be a bug in the atapicam driver: it doesn't provide these statistics. After several minutes, depending on the size of the track, the messages start to appear.

The Triébert bassoon is pretty much like the Buffet-Crampon instrument that I bought some years ago—in fact, it seems almost exactly 8 years ago, on 17 March 2000. Since buying the Buffet, I have done almost nothing with it, not even oiling it, of which it is in great need. Spent some time this afternoon doing that.

Darah's Greasy Heel has lived up to our fears. It's now been nearly 6 weeks, and it still hasn't healed. Yvonne and Chris into town today, also taking Samba for a pregnancy test. That proved worthwhile: Kurt Enzinger, the vet, tried a different approach and removed the scabs almost immediately. Let's hope that they stay away. Also, yes, Samba is pregnant.

Into town myself later on for a dental check and clean, the first time at the Eureka Dental Centre. He took jaw X-rays as a matter of routine, but without warning me of the price. Still, the results were interesting; they showed the problems I had had last year, where the periodontist had measured them with a probe.

Also looked in at the local home-brew shop to look for some darker crystal malt, but they only had one kind, and he wasn't too sure how dark it was. Bought some bread-making supplies instead; maybe we can find a way to get something better than the sponges that everybody here seems to prefer.

Waking up in the middle of the night with UPSs beeping because of a power failure is bad enough, but today it happened with a slight difference: a UPS was beeping without any power failure. It seems that it's not feeling well, possibly because the batteries have had it—not surprising after 8 years. Is it worth repairing? To be investigated. Now I only have one UPS left over from Wantadilla.

Somehow I didn't get much done today, perhaps because I had intended to tidy up my office, which was in need of it. Instead looked at some of the issues needing cleaning up after Sunday's crash, and got myself sidetracked reformatting the diary pages for May and June 2004, which was more work than I expected. It'll certainly be quite a while before I get everything converted.

Yvonne spent the day fencing off the south paddock, which also involved Chris' help pulling down the branch of a wattle that had been damaged last year:

That was certainly not too early: we had found lots of kangaroo droppings in the garden earlier in the day, made up for by the disappearance of our last Grevillea Thelmannia brought from Wantadilla; they had had a go at them some months ago, and eaten most of them, but today they ate the last one. It's fun having kangaroos around, but I hope they now stay the other side of the fence.

Another day paving the way to hell with good intentions: I had once again intended to tidy up my office. The best I can say is that I did a couple of minutes' work and tidied away some books.

My script for replacing spaces in file names has evoked a certain amount of feedback. One person suggested:

works with bash 3.2.17

spaceout()
{
        for ((i = 1 ; i <= $# ; i++)); do
                FN = "$(eval "echo \"\$${i}\"")"
                SFN = "$(echo "${FN}" | sed 's: :-:g')"
                [ "${FN}" = "${SFN}" ] && continue
                mv "${FN}" "${SFN}"
        done
}

This looks completely flawed. It's overly complicated, and there are obvious syntactical errors. It certainly doesn't come close to working with bash 3.2.33.

Peter Jeremy came up with a more elegant solution:

# Replace spaces in names with -.
spaceout ()
{
    for i in "$@"; do
        mv "$i" $(echo "$i" | tr ' ' '-')
    done
}

As he says,

I'm surprised you fell into the 'for i in $*' trap (and '(...echo ...)|sh' seems excessively convoluted).

Yes, he's right. Mea culpa. The use of sed is straightforward: I've always had a mistrust of tr, and sed is more flexible. The $@ is more surprising: I've never heard of it before in a shell context. It's been there for ever, so this was just a gap in my education. Without quotes, it returns a single entity consisting of all the supplied parameters. With quotes, however, it supplies individual words containing the original parameters, just what I needed. Thanks to Peter for opening my eyes.

That was certainly not too early: we had found lots of kangaroo droppings in the garden earlier in the day, made up for by the disappearance of our last Grevillea Thelmannia brought from Wantadilla; they had had a go at them some months ago, and eaten most of them, but today they ate the last one. It's fun having kangaroos around, but I hope they now stay the other side of the fence.

Peter had a number of other comments: the yellow flower that I mentioned a while back is a gazania, and there's another program I could try for cropping photos: netpbm. More stuff to keep me from tidying up my office, I suppose. At any rate, netpbm fulfils my expectations for descriptions:

About Netpbm

Netpbm is a toolkit for manipulation of graphic images, ...

The good news: yes, I got a little further with tidying things up. Also tried more bassoon practice.

From top to bottom, these are the butts of the my Buffet Crampon of uncertain age between about 1870 and 1930, an undated Hawkes and Sons instrument that I'd put at about 1930, the 1879 Triébert and the 1842 Savary jeune. The bottom three have the plain G hole, and in fact are almost identical in layout with exception of the rings on the Triébert, which I have never seen on another instrument. On the bottom three instruments, from left to right, are the three finger holes (right 1, right 2 and right 3, the A♭ key (normally closed) and the F key (normally open). The B♭ keys have been removed to show the hole, but they're in the same place as on the Buffet, which shows the plate for the G hole, the B♭ key and only two finger holes.

It's clear that the B♭ hole on the Triébert is higher and smaller than on the other two, roughly in the same position as on the Buffet. But what's the purpose of the rings? I've never seen any mention or photo of this particular arrangement. They would appear to be there to vent the B, but they don't help with the B♭.

So, I'm puzzled. This instrument wasn't cheap at the time. And it seems that the problems with the intonation of the keyed B♭ are due to the way it was built. Why? Triébert was known as an experimenter. Was this an experimental model?

I had intended to brew today, but went riding instead, down towards Corindhap. That wasn't long—somehow we never seem to go for more than an hour—but it showed us the limits of our little forest, and left me tired enough that I didn't do much for the rest of the day.

Somehow we always have too much food. I had bought some cauliflower and some beans for various Indian dishes, but Yvonne has had enough Indian food for the while, so she ended up imitating a dish on the cover of the Time-Life “Vegetables” cook book (but, strangely, without a recipe inside). The result was very close to the original, maybe exactly because there was no recipe:

I still think that the Indians do a better job of cooking vegetables.

Finally it's raining!

Finally got down to doing some of my “to do” list, and ended up spending hours on the phone. Also some bassoon practice—this exchange of instruments has had a positive effect in that respect.

Brew day today, and an absolute catastrophe. I had the mother of all stuck sparges, and I don't even know why: I brewed in the same way as always. Spent the entire afternoon trying to salvage what I could, and wasn't able to do anything else.

A friend in Italy has been trying to transfer money to me with PayPal for some time now, without success: the web site refuses to perform the transfer. He sent them a message and got the usual boilerplate reply:

Siamo lieti di soddisfare la tua richiesta di informazioni riguardo al pagamento che stai cercando di effettuare.

Se ti riferisci alla transazione di xxx euro che stai cercando di effettuare, questa é stata bloccata automaticamente dal sistema.

PayPal adotta alcune misure allo scopo di garantire la sicurezza delle transazioni online. Tali misure possono interessare alcuni conti in regola, ci scusiamo per gli eventuali disagi arrecati agli utenti.

Le procedure adottate da Paypal sono simili a quelle utilizzate dalle banche e dalle società emittenti di carte di credito e consentono di identificare schemi di comportamento durante il processo di approvazione delle transazioni. Il sistema adotta sofisticati modelli statistici per esaminare i parametri della singola transazione e se questi corrispondono ad uno schema di comportamento non ordinario, la transazione può essere automaticamente rifiutata.

Per garantire l'integrità della propria rete di pagamenti, PayPal blocca l'utilizzo di una specifica fonte di finanziamento utilizzata per la transazione che il sistema ha valutato come soggetta ad un potenziale livello di rischio.

About the only thing that my limited Italian detects is that this is boilerplate and doesn't explain the real reason they've blocked the transaction, and that they're using the familiar form of address, which seems out of place here. Spent some time checking with ANZ and discovered that the costs for an international transfer weren't so high after all, only $12, or about €7. Still, it would be better to save that much money. But my opinion of PayPal is further diminishing.

So finally my Italian friend got a friend to transfer the money via PayPal; at least it shows that the problem's not at my end. But it cost € 17.35! That's 2.5 times as much as an international bank transfer. Yes, the money was there immediately, but I'm left wondering whether PayPal is of any other use. If the banks got into the act, they could certainly offer PayPal a lot of well-deserved competition.

Spent most of a second day recovering my stuck sparge. It wasn't just the messing around: the stove in our kitchen is so weak that it took 2½ hours to bring a kettle full of wort to the boil, and I had to do it twice, the second time mercifully only with a half kettle. Time to look round for a better stove, not to mention bigger kettles.

The events of the last couple of days don't make for exciting reading, but they certainly took it out of me. Spent most of the day relaxing, watching TV and doing some bassoon practice. It's becoming increasingly clear that the reeds I have aren't worth much.

Today was my father's 85th birthday, so off to Maffra to visit him. Four hours there, four hours back, and 2½ hours in Maffra. Still, it was worth it—it's not clear how much longer he'll be around.

Freda came along as well, and we spent far too much time talking about the dim and distant past. Even our “recent” family reunion was over 30 years ago.

Another quiet day. I need to take another stab at tidying things up, but today wasn't the day. Spent some time looking for camera lenses and other stuff on eBay. That takes a surprising amount of time.

More bassoon practice. For some reason I can't get high notes (above about a above middle c) on any of my old French bassoons. I suspect it's related to the reeds and crooks. Where can I get good reeds? Do I need to learn to make them myself after all?

Yvonne had a dentist's appointment today in Geelong, so went with her to take a look at the town. While she was at the dentist, I went to the RACV shop to buy some travel books. Five minutes looking for books, and fully ten minutes waiting to pay—RACV have this crazy lack of priorities, so I had to wait while all available cashiers dealt with customers requiring advice. It seems that they were short-staffed, but that's no excuse for that kind of treatment of customers who just want to pay.

In the end, decided to take a look round the remaining parts of the Great Ocean Road that I had not yet seen, from Geelong to Apollo Bay. Somehow I had a feeling of déjà vu driving along the road, probably because I really had seen it from so many photos.

In Lorne headed for Erskine Falls, which were conveniently close to the car park. A good thing too: you shouldn't visit waterfalls in a drought, and there wasn't much to see:

Decided not to continue to Apollo Bay and cut across country towards Forrest, saving an impressive 50 km in the process. We had planned to visit the Colac Botanic Gardens, but time was getting on, so put that off for another day.

Do you have a comment about something I have written? This is a diary, not a “blog”, and there is deliberately no provision for directly adding comments. It's also not a vehicle for third-party content. But I welcome feedback and try to reply to all messages I receive. See the diary overview for more details. If you do send me a message relating to something I have written, please indicate whether you'd prefer me not to mention your name. Otherwise I'll assume that it's OK to do so.