Linked by Kaj de Vos on Tue 8th Jun 2010 22:07 UTC
Syllable, AtheOS

The Syllable project is pleased to announce that the reports of its death have been greatly exaggerated, and that the new version 0.4 of Syllable Server has been released. This release focuses on maturing existing functionality, improving security, ongoing system restructuring, and making the system a suitable base for third-party package managers. Although the project admits it hasn't brought its unicorn factory online yet, extensive work was done on the nitty-gritty, so the full change log is quite long.

Permalink for comment 429348
To read all comments associated with this story, please click here.
RE[13]: Cool, I guess...
by Kaj-de-Vos on Thu 10th Jun 2010 00:50 UTC in reply to "RE[12]: Cool, I guess..."
Member since:

"In the case of Syllable, by "everything" you mean that less than one percent of our own code that is in Syllable Desktop?

I don't really understand why you're so proud of having not so much new code, in itself this doesn't mean much as it may have very different interpretations:
- positive: you reuse much code from other OS
- negative: you don't support much hardware / your applications are straight ports from other OS, which means that if they suck on these other OS, they'll suck equally on Syllable Desktop.
I'm proud of Syllable. Why should I be less proud if the code that constitutes it turns out to have a different makeup?

I don't know if you're a glass-half-empty guy, but the fact is not that we have little own code, but that we have a lot of reused code. I'm also proud of that, because we add value to that existing code, and this is the ideal thing to do in an open source ecosystem.

Note however, that I'm not defending any specific interpretation, just saying that this number alone doesn't mean much..

To the contrary, it is essential. It is essential to the discussions about what we should and shouldn't do, because it determines the associated cost balance. We get a lot of flak for doing Syllable Server from people who are convinced that it damages Syllable Desktop. This code analysis proved long ago that this is a fallacy. Even if it is hard to believe, Server does not dilute the technical quality of Desktop and does not take resources away from it, but rather helps it, because so much of the effort is shared.

It also points to the psychological effect that's at work here. Syllable's reason to exist is that we have always maintained that Linux/X11 is not good enough and is doing fundamental things very wrong. Linux itself has been characterised as a reactive project, and in turn, many Syllable fans are strongly opposed to Linux. The fact that Kurt has achieved so much better performance with AtheOS while creating only a few percent of his own code proves in a big way that he and we were correct. We have been extremely good at replacing exactly the parts that matter. Both camps hate to hear this. The Linux/X11 camp obviously doesn't want to be proved wrong, and certainly not in such a shattering way. It threatens their self-created hero image. But Linux-hating Syllable fans don't want to hear it, either, because the proof says that Syllable is 99% the same as their self-chosen object of hatred, so it tosses them into an identity crisis, and thus, denial.

So excuse us if we don't care much for either camp, but draw up our own plan of action.

As an aside: congratulations for having chosen a port of WebKit instead of Firefox for the web browser, is-it multi-process on Syllable Desktop like Chrome(*) is?

* I've switched recently from Firefox to Chrome, and I really enjoy Chrome's responsiveness.

Thanks. Our current WebKit port in Webster is from before the multi-process feature, as far as I remember. From Haiku, I understand it's also very system-specific to implement. However, multi-process is not about responsiveness but about stability. Chrome is responsive because it uses WebKit instead of Gecko, but mostly because it has its own toolkit that has a very BeOS-like design. Syllable's native toolkit is like that, too, of course, so we have always had the toolkit and the engine to achieve similar results. If you open multiple Webster windows instead of tabs (which is the fundamentally correct way, anyway), you also have multi-process.

Reply Parent Score: 1