Older blog entries for dto (starting at number 36)

Another rant from Dave.

Yes, it's that time of the week again. Looking at stuff like this makes me get just a tiny bit upset about how badly the linux world is dragging its political feet with respect to improving the interface. I'm not talking about making all the OK buttons respond to the Enter key (currently my biggest pet peeve about GNOME, and it's slowly being fixed---recent GIMP etc.)

I'm talking about the imaging model. I don't want to criticize X unfairly. The X Window System was brilliant for its time and in its environment. But it simply does not support what people want to do now well enough to continue. Fast vector imaging, transparency, high-resolution monitors, antialiasing. Yes, you can implement software on top but there's no standard and it's slow.

The first defense I hear all the time is network transparency. I respond: who cares. Most people who actually attempt to run a remote X application run away screaming, because if you don't have a T3 it is just unusably slow. Sysadmins who actually do need remote execution just use ssh or telnet anyway. And if you absolutely want/need remote X execution, there is nothing stopping you from running an X session on whatever box you wish to use. (Presumably if you have the resources to make remote X execution usable, you will have a spare box for this.) Or running an X session on top of whatever succeeds it. Even if this software is hard to set up I cannot imagine how it could be more complicated than configuring X in many situations.

I don't mind if the New World Order is just X12. Some kind of fresh start is needed. If linux is to achieve the final unification of UNIX, we must be able to break with the past when it no longer suits us. Get rid of vi? Nope, your choice of text editor affects only you. Get rid of bash? Same here. But X is a platform, and it's no longer filling the requirements of a modern graphics-oriented user OS.

As for the anti-chrome folks, they do have several points about not making hideous bloaty bitmapped skins for everything. But there's more to an imaging model than interface widgets. There's no really good vector imaging program available on native linux. That's because there's no good vector imaging model. There will be little print/design success for linux until this happens---GIMP is a fantastic program but both bitmap and vector software are needed.

We also need real file management. Even if nautilus ends up being twice as good as Windows Explorer, it won't be half as good as OS/2's Workplace Shell. Anyone from IBM reading this? I hope I don't sound like an Amiga Workbench Acolyte, but the OS/2 interface was pretty nice in a functional sense (although visually bland, which we can fix.)

It's frustrating. I don't want to become one of the "don't imitate windows!" crowd, but even without innovating at all we could do so much better! Well- understood paradigms like OS/2 WPS and Display PostScript (please help us Raph!) are out there and probably well- documented.

Believe me, I am a "worse is better" person. I'm into the UNIX philosophy. But there comes a time when we simply need a new piece of software because the old one, however loved, isn't getting the job done. I think we are there.

I finished (for the time being) my Antarktika page. This comes in the wake of the recent announcement on idmlist that I will be doing a release and shows with Systorm Technologies in the coming months.

The recent Sawfish updates that were put into helix-update are BAD! They will crash everything. If this happens to you, you must delete ~/.sawfish (perhaps backing up your themes dir first, as I forgot to do.) Then restart X and things should work ok.

I didn't get much accomplished this weekend... lots of reading, but not much real work. Tonight I'm planning to rewrite/finish that little PHP backend I've been working on for Igloo.

8 Oct 2000 (updated 8 Oct 2000 at 19:10 UTC) »
<rant&rt;

I think it's important that everyone take a look at the ACM's official position on the licensing of software engineers:

ACM Software Engineering + Licensing

I think the politicization of computer science and software engineering is getting worse. After decades of spotty success in having the academics come up with practices that work in the real world, there is now the chance that they will be mandated instead.

The opportunities for corruption and exclusionism are unbelieveable. I may sound like a conspiracy theorist, but look at the ACM task force's position:

"Licensing (particularly that applied only to a small group within a profession, i.e., those selling their services or consultants) has in other professions been used to keep people out and to create guilds controlled by those who stand to profit from them."

The destructive power-group potential of licensing is explicitly acknowledged in the report. The software engineering world does contain a core of solid practice, but its public face (and its money) is dominated by counterproductive and overpraised fads. The primary purpose of it is to make money doing consulting and staff training---witness the complexity of UML and all the expensive books and tools. None of it seriously works, except for the creators of course. Imagine all this becoming mandatory---imagine the software industry being forced to follow the dictates of who's got the committee's ear (or wallet) instead of what works or what is needed.

I'm not saying that certification could have no benefits. But doing it at this point, when the sediment of good practice has only begun to truly settle and harden, would be a disaster. </b>

7 Oct 2000 (updated 7 Oct 2000 at 23:16 UTC) »

Just got back from a great bike ride in Shrewsbury with Michal. It's been a pretty decent day... Ari's not around, but I don't have any schoolwork to do and it's a three-day weekend!

For some reason the #gnome channel I normally hang out in has become "invite only". I can't get in. :-(

I'm going to implement the Octal mixer opcodes tonight, and probably test out the finalized unit generator network now that the V2 engine code is done. Perhaps tomorrow I'll add more of the multithread mechanism (dispatching calls.)

6 Oct 2000 (updated 6 Oct 2000 at 05:34 UTC) »

Eeh. Sounds like Michal is pretty stressed at the new school.

The grade situation is fixed. Unfortunately the User Interfaces class will probably not complete---it's effectively cancelled.

Got a lot done today---I'm trying to clear up my weekend to get a huge chunk of work done on Octal. We have Columbus Day off, so I'll have three full days.

Rant mode! And frankly, Octal Must Get Done. NOW. I've got a really solid and flexible API going, one that covers issues that the LADSPA group is still arguing about. This has been well-reviewed by the coders on the octal mailing list.

I don't think there's too much really wrong with LADSPA itself, other than not doing enough. I just think they're going about it in a somewhat futile way. In userland, cool apps drive the adoption of new standards, hardly ever the other way round. Even the best one will fail to capture mindshare if no cool apps use it.

A cross-app audio API is the solution to a problem that only exists if we have stellar audio apps, which we don't. We're getting there, but to be honest there is nothing remotely as good as Buzz for music production available on this platform.

Whatever API the "killer app" uses, when it comes, will be the de facto standard, no matter what the storybooks say. If Octal achieves even half of the kind of community enthusiasm that Buzz did on Windows, even half the market penetration, then the GNU OX_API standard will succeed. I feel there's a good chance of this happening---it's small, has very few dependencies, has more features than Buzz... the "gimp of audio" may finally happen, as a suite of apps like GLAME and OCTAL.

And then the rest will fall into line and we will have plugin sharing. I say this not in the spirit of "world domination" but because in my opinion, this "tail without a dog" situation is just resulting in many projects adding minor support for LADSPA so they can be "with it" but without any of them getting finished enough to gain any real userbase so as to spread the standard. Even worse, because the standard was designed to be integrated into so many existing projects that used their own API's already, it is just too minimal. It's probably easy to implement, but that's because it doesn't do a whole lot.

User interface issues (widget types, etc) are underdeveloped, there's no multitracking capability or the direct voice control policy as tracker fans enjoy, there's no memory allocator wrapper for use by plugins, there's no standard data types (with concomitant standardized visual/textual representation) to be shared across different plugins, there's no built-in host callback extension mechanism like in OXAPI, there's no built-in musical utility functions, and there's no wavetable (and in the words of their newsgroup, "no good way" to add one). There's really not much beyond some function pointers, typedefs, and a few suggestions.

And, since its approach to control data feels very "visual CSound", we'll probably end up with a bloated interface like most of the linux "softsynths." The classic "visual programming" idea, where every little box on the screen is festooned with numerous i/o ports that can be arbitrarily interconnected without rhyme or reason. It's The Interface That Won't Go Away. Nobody wants these 100 little boxes with tons of 4-pixel-wide labeled control ports. Unit generators are already complex and expressive enough without the tinkertoy UI that people keep trying to do. The reason for Buzz's success above all the others is the simplification of the unit generator UI but not much loss of power.

The CSound kind of stuff (with the exception of judicious sample creation) isn't widely used to construct real music. It's the kind of Max / kyma stuff that leads people into "digital fuckery" where there's no music but tons of cool freaky sounds and ultra-precise rhythms. This is a rant but the technology is definitely a factor in the current uselessness of "modern" idm, where the only good new stuff sounds like it was made in 1994.

Rant mode over :-)

While watching the Vice Pres. debate, I finished off my Principles Of Programming Languages assignment. Now I've got to study for tomorrow's exams. Then it's free for the long weekend.

It rained today, so no bike ride. Awww :-(

Whoo! I am just a regular powerhouse. Ariane and I have been biking together every day. Yesterday I caught up in Physics. Today I completed my second Assembly Language assignment and wrote a paper on IP headers for my datacomm class.

Unfortunately I had to use Word, because for some reason my instructor required it. No problem... I typed and edited it with vi, and then converted to word with one of the windows boxes downstairs. The file size went up by a factor of 20.

I briefly talked to miguel the other day about Gnome and sound apps, and how I might make a contribution there with GNU Octal. Should be an interesting line of discussion, sometime when he's not busy :-).

It looks like my professor will be out for the whole semester due to her illness. Now I have to scramble to find something that fills the elective credits I need before graduating. This situation has potentially screwed over me and a lot of other CS seniors who have little room for error.

random thought: I was initially fascinated with wikis the first time I ran across them a few years ago, but after more experience I've come to the conclusion that they're too diffuse and they tend toward "mental masturbation."

Octal is about 65% finished.... I've got to get on the ball with this one. I've got to finish it. I'm depressed. Ari's not around. Michal is too busy. I have some new fall clothes but now it's warming up. :-(

I'm behind in Physics. Not only that, I missed a lab. This Sucks. I'm very good at catching up, but it's so damn stressful...

wow isn't this a bitchfest?

28 Sep 2000 (updated 28 Sep 2000 at 00:59 UTC) »

EBay sucks for games. I've lost several auctions at the last minute to Big Shot Collector types. I figured I'd get something. I'm not a professional collector. I just want to play and rebuild my library of games. But the vultures who run a business collecting and selling retrogames jack up the price and snatch it all away from people who just want to play. Soon some fat hairy guy in Minnesota is going to own all the Intv gear and I'll never have a replacement.

The reserve is usually something like $10, but by the time the auction ends it's been jacked up to $60 or even way more. This sucks. I would not pay that much for a whole lot at a garage sale.

My new quest is to obtain an Intellivision II with System Changer. I already have a working Intellivoice, but the system itself is slowly dying. This single sleek unit will play the best home games of the 70's and 80's (not counting Metroid.)

God, life sucks.

27 older entries...

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!