Older blog entries for stefan (starting at number 10)

lilo seems highly confused after an excellent diary entry from graydon

Noone is suggesting that we watch more TV. And the mere existence of web sites is of course no problem either. It's the rate at which stuff proliferates, be it web pages or cars. This is growth getting out of control, it's a cancerous tumor.
Each one of us is consuming a lot, and of course finds good reasons (a posteriori !) to defend that fact. A healthy approach would be to step back, to distanciate yourself from yourself and your surrounding. I guess that's what self consciousness is all about.

As to needing 'pretty dictators' as excuse for not being able to feed the whole planet, that's plain rediculous. You mean, it is the fault of all the Mugabes, Ghaddafis, or Husseins that there is such a huge gradient in wealth between what we refer to as first and third world ?
As graydon points out, it is our inability to reason, the all so modern religion of 'free markets' which is at the heart of all that, both, the world's poverty, as well as the uncontrolled growth we are experiencing.
Isn't it strange, how keen we are on technological excellence, controlling even the genetic foundation of future generations of living beings, but at the same time throwing up our hands if it comes to having the slightest control over us as a society / culture ?

Raph: thanks for the link to I Don't Like Your Examples! . I found this article most amusing. How true ! Even the most unrelated messages in whatever media carry some hidden message, remembering you gently of the limits of political correctness. This is the modern form of totalitarian control. Totalitarian, because it's universal, ubiquitous. You can't escape. Like the daily (nightly) messages in the dormitories in "Brave New World".

raph writes:

My heart is telling me that, while I could upgrade the X driver for Ghostscript to do shared memory, Render, and so on, I'd rather share that work with a more generally useful framework than do something completely Ghostscript specific. The framework that has what we need now and is a good base for future work is Gnome.

I'm still not convinced that any GUI or window system integration is needed here at all. (What else would you expect from me ? :).
There are two aspects to be considered:

The renderer can be separated into a low level driver, which cares about buffer layout and h/w acceleration, and a high level PS/PDF interpreter which talks to the driver. Have a look at the GGI project to see how this can be achieved. GGI ships with a bunch of output 'target' drivers, providing GGI access to X, /dev/fb, (shared) memory, svga, and many more.
GGI applications therefor are intrinsically portable.

The User Interface with which you control the ghostscript process, should be a simple frontend, i.e. should not be part of the project itself. If you provide a sufficiently complete API providing access to the controlling attributes, someone can attach his own graphical front end to it without problems. I don't see where this would cause any restriction on ghostscript. On the contrary !

Coincidentally, someone wrote a little (naturally poor) ghostscript-for-ggi tool, which uses a pixmap to transfer the rendered image to the ggi process. I proposed to write a ggi driver for ghostscript instead.
In fact, it would be an extremely useful thing, since berlin would immediately have a serious application.

On a related note: my recent experiments with shm Drawables in Berlin are all encouraging. I hope to be able to provide the means for emulation libraries (Xlib, gdk, etc.) as well as high bandwidth video access. Finally people can start porting programs to berlin. Horray !

aaronl write:

Component systems blur the boundries of what is one application and what is another. This would be nice if it could be put to good use. But the fact is that the only thing it is good for is bloat. Once an application supports Bonobo or a similar library, it can do anything.

...which is exactly its purpose. UI designers have long asked for computers to adopt more to human needs instead of requiring us to think in computer terms. This implies that we think of documents and their parts and not in terms of applications which can deal with them.

Of course you can abuse such an architecture. But that's not the system's fault. It's remarkable that you speak of Emacs as a counter example (or did I misinterpret that ?). Emacs is - in this respect - for me the worst scenario: you slam everythin into the one big application. *That* is what I call bloat.

However, I don't want to browse the web with my file manager (if ever I use one), be it componentized or not. The point is about composing a humane interface with the means we have around. It seems to be tempting to abuse the power of components, 'because we can'. May be as components become more the norm we will make better and more effective use of them.

I'm impressed how a simple remark evolved into quite a thread with some very eloquent statements. Having participated in quite a number of similar discussions, I'm still puzzled about what can be discussed and what is by definition impossible to argue about.

My (non-provable) observation has been that it's always matter which is at the origin of everything else and motion (evolution) is part of its nature.. Of course, to give some substance to this claim I'd need to give a precise definition of the term 'matter', which I won't. Only one thing: it's much more than just particles, or fields, or whatever is the object of interest to, say, physicists. Matter in this philosophical context is anything which exists outside of my reflecting mind, i.e. especially the conditions of my life, i.e. whatever makes me think now the thought I'm typing here...

This is nothing I can really argue about, it just fits my experience. On the other hand, there are quite a number of things I can derive from that position, which then are coherent and defensable. All these categories as beauty, aesthetics, value, moral, etc. are purely anthropomorphic. As lilo illustrates, these categories can't be applied outside our activity horizon. Well, they actually are, but I think that doesn't make sense. In particular, the notion of 'sense' has meaning only in the context of 'consciousness' and conscious choice, since there we give sense by chosing one option over another.

To call something 'logical' merely states that our (logical) thinking is well able to represent that thing, it doesn't make reality more real because we agree. Nature doesn't care about logic. We do.

What I find even more interesting in this context is the question why we think and feel what we think and how this can be derived from our culture, the way we have been educated, the way we engage into any kind of humane activity, etc.
But lo!, here we have our famous catch 22 ! That our thinking is a function of our social condition was an assumption. What if people don't follow me this far ? How can this give a little substance to a meaningful discussion ?

integrated openC++ into synopsis as a C++ parser frontend. OpenC++ is an awesome tool, it actually understands real C++ grammar. So we now can parse anything from <signal.h> over <string> to <CosNaming.idl> without cheating. Synopsis proves its usefulness as I try to integrate an IDL manual with a C++ manual. Hopefully that will encourage people to play with berlin as this is the ultimate reason I started the synopsis project...

While I'm reading the OpenDoc specs / reference and John Vlissides' Dissertation about "Generalized Graphical Object Editing" (describing the Unidraw framework which was developed as part of InterViews) my head starts to spin. It is a real pain if you have to figure out how things historically relate to each other just because there is close to no other documentation available. Anyway, things start to make sense again.
I'm now wondering how to take all that and create a Compound Document Architecture for berlin with it.
I'm also glad Michi Henning confirms that a particular POA based design I intend to implement for berlin's kits is precisely what POAs were intended for in the first place. No more worries about distributed garbage collection. Phew.

Mathieu: I just read your OAF documentation and I wonder why you don't simply use the (standardized !) trader interfaces (plus some equivalent of an IMR of course). I had a similar discussion with Miguel about monikers vs. IORs. Both issues make me a bit nervous. Of course, I'd like you to use standardized CORBA services/facilities because this would make a cooperation between bonobo and berlin *much* easier...

after OLS and a holiday week with family and friends, I feel quite productive this week. I got synopsis to produce a nice Reference Manual for berlin (which you can have a look at here), and I finished porting berlin to the POA architecture, which was amazingly simple. Now I'm reworking the Console layer in berlin to permit various packages to be used (GGI, SDL or - as I discussed with NickElm - even exotic things like a console for the cube. Let's see whether we can get there...

omniORB 3.0 was released on friday. This is big news for berlin since we now can port it to a POA based architecture and with that make it completely ORB agnostic.
by friday night berlin already compiled and run with the new omniORB, by now it compiles even with POAs, but it crashes in the plugin loading process. That's not very surprising given the whole lot changes. It's easy to imagine that I got one of these soubtle issues for example about reference counting and resource (de-)allocation wrong. Some exciting debug sessions are awaiting me, once back from the OLS :)

free software and economics

If Free Software and its consequences had only one impact on society, it was this: to catalyze discussions about how to design a better society. Lots of these topics are still hugely put under taboo or otherwise dogmatically treated. Often I find comments like this one from a recent diary entry:

Capitalism is an economic model designed (in broad) to maximize value to consumers

I don't think that capitalism was designed at all (and certainly not for the consumers). Quite on the contrary, some modern positions are to keep out of 'regulations' and let nature (in form of free market) act itself. They seem to confuse conscious design with authoritarian regulation and imposition.
I think making this step towards consciously designing our society is as much an evolutionary and emancipatory step as is the transition towards Free Software.

Raph: much the same way I don't think that there is any creation of ' "artificial" markets'. It is in the nature of capitalism to commoditize everything. If the 'right to pollute' is treated as a commodity this is not because someone thinks the market is best at optimizing this new 'resource', it is because it (unfortunately) can be treated that way.
It is similar with Free Software: we stipulate new rules concerning the distribution of software, and we watch how new business models emerge together with new commodities around this new entity.

The question is whether this can be sustainable and stable. As much as I admire the good will of people to think about alternative models, I still find the restriction to free market scenarios too limiting. At least in the sense mentioned above, i.e. the categorical denial of any design, as if men were by nature unable to take there fortune into their own hands.
Society has always evolved, I don't see any reason why this time it should not progress. Eben Moglen has some very good points to make in his article. Let's be optimistic...

back from Stuttgart's LinuxTag event which was truely amazing. Beside getting in touch with a couple of other berlin developers such as NickElm, I met Thomas Hiller, the ancient Fresco maintainer. It is sad how much work was done on Fresco and is lost simply because at the time no-one cared.
Spend a couple of days with friends in Z├╝rich after that. It's fascinating to hear the mix of French/German my daughter speaks now, having spent almost three weeks in Germany/Switzerland.


I think the problem with such discussions about freedom is that people forget that it's not meaningful to speak about freedom without addressing to whom it applies. Saying that a piece of software is free doesn't tell anything as long as you don't say whether you mean to the developer, user, or vendor.
People are of course right if they claim that a BSD style license is more free than GPL if they mean for the vendor. But since this freedom is (potentially) restricting the user's freedom, and since the user is what the GPL is addressing, there is a deep conflict.
I think as soon as you add this little bit of context into the discussion it isn't any more about which license is more free but which interest you are trying to defend.

1 older entry...

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!