...and Then Some.
Update: Davi wanted to take his occasional pot shot at me (it was about time, I guess), so I added some background where relevant. He likes to make it personal, I like to make it factual. As always, you are free to make up your own mind.
Sometimes, I just have to vent some steam. And not just where it regards news commentary, either.
Brain-dead Windows PowerPoint Dithering
If you've ever tried using PowerPoint on Terminal Services or Remote Desktop, you probably noticed that even at ludicrous bit depths (16 or 24 bits), PowerPoint 2003 persists in using idiotic GDI dithering for its display unless you're either in slideshow or animation preview mode.
In standard editing mode, however, editing anything against a gradient backdrop is completely impossible. And yes, I have to do this on Windows - only Office 2003 can deal with the embedded data on some of the files without mangling it.
There is no known fix. I asked someone at the Office team a few months back, and she never even heard of the problem. I spent the last three days running Office remotely from a Linux machine and trying to get around it, even down to going through the rdesktop source. As far as I can understand it, PowerPoint tries to use hardware acceleration, and somehow the detection logic is different when in editing mode - so it falls back to GDI dithering.
Linux Wireless Networking
Even on Fedora Core 4, it is still a mess. I cannot believe that I still have to re-enter WEP keys manually and run arcane commands to do something as simple as switch wireless networks (or get things to work properly when I plug in a cable).
Yes, yes, I know that Ubuntu and SuSE have somewhat better wireless management. I couldn't care less - it's not about each distribution's little nerdy tweaks, it's about the whole thing - there is no common way to do this across distributions (and believe me, I've seen a few).
And don't get me started on power management, which was the final straw as far as my home laptop was concerned - when my iBook died, I turned my ailing Toshiba into a VNC-based "network computer" - I tried tweaking Linux, but using XP I can get it to last for nearly three hours on a full charge, and it hybernates properly.
Bringing up the Activity Viewer window hints at some sort of interlocking between several parallel mailbox accesses. I can understand that other "Opening Mailbox..." threads to the same server might have some impact (although the server runs smooth as clockwork), but I'll never understand why stopping tasks accessing different servers makes things work.
After half an hour of waiting for Mail.app to try to move those messages, I found that clicking on my inbox folders showed no messages at all. Mail.app went catatonic, and I just switched to Thunderbird.
It moved everything instantly.
I mean, client support is pretty damn bad as it is, but why can't Apple's own server do this properly? OSXVnc gets the RFB optimizations mostly right (it outperforms Apple's drastically), but none of them can deal with Windows or Linux clients using a Portuguese keyboard.
It's not as if keycode handling is rocket science. Or, better still, that there aren't droves of non-English speaking Mac users out there willing to pay $25 for a client and server that does all of this properly (or even a bit more for a Windows-compatible RDP server).
And no, I don't have the time to do this myself, even with my usual 5-hour-a-day sleep cycle. The whole point is that I shouldn't have to fix this sort of thing, not when Apple includes it as part of their own Remote Desktop solution.
Proprietary Phone Connectors
I've ranted about this before, but this week I had the occasion of dealing with yet another connector change from a major manufacturer, and it was so idiotic and underhandedly stupid that I will be doing the utmost not to mention it anywhere on my site, lest I go all out and employ some of the quaint (but exceedingly colorful) expressions that one of my Scottish colleagues used.
There is absolutely no sensible reason why phones need to have proprietary data or audio connectors - it simply can't be about mechanical resilience, shrinking margins or integration effort, not when USB chipsets are dirt cheap and can present just about any internal interface manufacturers might need.
Trust me, I should know. I switch phones around three to seven times a month, and the only thing that I stick to is my Blackberry - which charges, syncs and is firmware upgradeable from a standard mini-USB connector.
That said, there must be a special place in Hell for whoever designs USB cables with proprietary plugs at the phone end. And I'm not talking about "sub mini" connectors, either.