It's Massive Disappointment

That's how Pierre Igot describes OS X 10.5.3.

Get It Try It

Pierre Igot finds himself baffled. He's baffled Apple take so long to produce a very much needed update - and then update so little. He's particularly worried (as others) about the literal swarm of bugs that never get fixed and seem to be summarily ignored.

All the while the dweebs add a zillion gajillion Automator actions.

Nova Scotian Pierre has run the Betalogue blog since 2003 and has recently written quite a few articles about his dissatisfaction with his computer vendor of choice.

Pierre's greatest alarm has been the dysfunctional desktop search system Spotlight which he claims can be directly dangerous to use.

Pierre calls Apple's recent 10.5.3 update a 'massive disappointment' in that it's both massive and a disappointment.

I don't mean that it is massively disappointing - only that it is quite massive (536 MB for the combo updater) and that it is also quite disappointing especially after such a long wait. I suspect those who do like to use Leopard's Spaces will appreciate the fixes included in 10.5.3 but for the rest of us out there the number of bug fixes that actually address the substantial usability issues with 10.5 is far too small.

This of course from the company Eric Raymond predicted could win the battle for the 64-bit beachhead - as far out in front as they were at the time.

Baffled

Like many others Pierre is baffled Apple don't fix their ever growing swarm of bugs and after describing a bug capable of breaking an entire WordPress installation fells the following.

How can Apple's engineers live with the fact that this bug still exists in 10.5.3? It has been a known issue for many weeks now and they have had more than enough time to address it. And this is not just a cosmetic bug users can live with - it is a potentially very destructive bug!

I cannot believe Apple's engineers don't realise this and don't make fixing this bug a high priority.

Pierre submitted a bug report nearly two months ago and considering how scary the bug is naturally expected a fix with the new update 10.5.3.

Just a few days ago I was looking for a specific keyword in a bunch of text files scattered within several subfolders inside a main folder. So I selected the main folder and initiated a search for the keyword in question with Spotlight in the Finder.

I got a Spotlight search results window with a list of three results in list view: three files named A, B, and C. I selected A in the list and then double clicked on it to open it. After the file was opened I could still see the search results window in the background - and sure enough: the name I'd selected and double clicked on had suddenly become editable - in the background.

And that's of course a no-no which all user interfaces steer clear of - for obvious reasons. But that's not the end. This is Apple. Things get worse.

But what's worse is a few moments later I saw in that background window OS X arbitrarily changed the name of result C (which was not even selected) to the same name as result A!

Note Pierre is not saying the system changed the name in the search results pane - he's saying the system arbitrarily changed the name of a file on disk.

Since result A and result C were two files in two different subfolders I didn't get the error message complaining that I was trying to rename file C using a name that already existed.

OS X simply randomly changed the name of result C to the name of result A and if I had not noticed it right there and then I would have continued my work and never realised I now had a file whose name was now incorrect.

And never mind this is totally unheard of in professional programming - it's directly dangerous.

And this happened in a context where file names are particularly important since the search in question had to do with the PHP files inside my WordPress folder - to be blunt: in such a context a changed file name means the entire WordPress installation is broken.

Back to Snail Mail?

But Pierre's not yet getting started. An equally sensitive area is the increasingly deplorable state of Apple's mail client which has suffered a continuous degradation in quality since version 10.3 back in October 2003.

Another area of 10.5.3 that is utterly disappointing is the Mail application. As far as I can tell none of the new problems introduced in 10.5 that substantially affect usability have been fixed. The alert sounds for incoming and outgoing mail are still played randomly with neither rhyme nor reason. Mail still does not play these sounds through the correct audio channel. It still removes the text colouring when replying to a message that has been coloured by a rule. Etc. Etc.

Things Used to Work Just Fine

And what newcomers can't appreciate - what long time users more easily see - is the software used to run 'just fine'.

And what the more alert users like Pierre wonder is how - and with what twisted talent and dysfunctional corporate organisation - you take code that's worked perfectly for years and then just systematically ruin it.

The issues in question are not as potentially damaging as the bug with Spotlight search results windows but they still affect the usability of the application in very visible - and audible - ways. Given that these are not new features but things that worked just fine in 10.4 and that Apple broke in 10.5 you'd think they would demonstrate a little more zeal in fixing them so that at least Mail works as well in 10.5 as it used to in 10.4.

But no. Now 10.5.3 is out and none of these things are fixed. This means that we still have to live with them for at least another couple of months if not longer.

I am afraid this is really quite disappointing.

The Three Rules

Seeing as Apple traditionally deal in nonprofessional systems and have deliberately kept themselves isolated from the programming community at large it's understandable - if not excusable - that they never heard of the three golden rules of programming - much less ever came to within a light year of observing them.

But there are three golden rules everyone else in the industry has used all along with ostensible success and it might be an opportune moment to review them.

These three rules were proposed by Brian Kernighan, the Bell Labs luminary who was responsible more than anyone else for getting the Unix philosophy out the door in Murray Hill.

Keep it simple. When your marketing rep comes to you with another one of his whiz-bang ideas and you know by now how difficult it is for him to extract his cerebrum from his anal orifice long enough to think clearly for a fraction of a second you'll know instinctively you must whittle his latest and greatest down to something that actually can compute.

Good computing ideas - and good computing systems - remain essentially simple as simplicity is the goal of all good design. This in contrast to Apple who don't think much about systems but only about devices they can get to market in time, that look cute, are overpriced, and are guaranteed to start falling apart in two years.

Build it in stages. The very fact you have a desktop search technology - Spotlight - that somehow gets a notion and a carte blanche to start renaming files on disk when it's supposed to be a bloody read-only application in the first place is a sure sign something's very very wrong in Cupertino. This 'technology' wasn't built in stages - that's for sure.

The advantage to - the reason for - building things in stages is you're supposed to bloody test your stuff at each 'stage' - as you add to it. You're not supposed to find yourself half a year down the road tossed about in a sea of millions of bug reports. You're not supposed to let this situation get out of hand as Apple have.

Let the other guy do all the hard work. This is particularly important and is something Apple have never been able to do.

You don't go reinventing the wheel. If someone's already made a good kernel you can use you don't find ridiculous excuses to change it. You don't listen to the insufferable Grubers and Siracusae out there who want the old days with nonprofessional file systems back again - you just take the FreeBSD and other technologies as is and live with it.

Every time you assimilate a new technology by modifying it you spread yourself thinner. Apple today are spread about as thin as they can go. Almost no open source technologies in use in OS X are 'out of the box'. All this means additional - and continual - work that takes you away from those tasks that are really important.

And Apple have their iPod and above all their iPhone today to distract them even more - and as everybody knows: Steve Jobs doesn't believe in hiring on more programmers. That never worked for Microsoft.

Paul Ingraham sums it up best.

I was also acutely annoyed by the unimpressive list of repairs in this 10.5.3. I've struggled with numerous Mail bugs since 10.5. I was eagerly awaiting repairs and I didn't get a single one of about 20 I was hoping for. It was truly a bummer to scroll down that list, realising they somehow managed to fix nothing I cared about.

I wholeheartedly agree the lack of substantive fixes of prominent bugs is baffling.

I don't think 'too much coffee' or a bad attitude causes one to be baffled by such things. I will never accept that simply being 'better than Windows' is enough for the Mac: that's too low a bar to get over.

OS X could be a lot better than it is and there's not really any good excuse for Apple to fail to take it all the way.

Much of the most desperately needed bug fixing is held up primarily by unwise allocation of human resources, by corporate prioritisation that pulls too much talent out of OS X development and puts it into iPhone work. If there's a complexity problem interfering with making OS X truly good it's Apple's institutional complexity, not OS X's technological complexity.

Apple cares about good design more than most companies but still not enough for a truly friendly technological future.

See Also
Betalogue: A Massive Disappointment