Older blog entries for glyph (starting at number 4)

halcy0n: Have you considered using Python? and Twisted? I think it makes writing networked apps a lot easier than it might be otherwise, but of course I'm biased :-).

I could say more about Perl here, but others have said it better. (Or should I say "bitter"?)

Crazy

This is an idea I've had for a while, but I have been tinkering with a bit lately due to some work stuff. I want to write a wiki which is just a regular website, on a filesystem someplace, that you edit with Mozilla Composer.

(Yes, I'm aware that XSS is a concern here. Partially that's why I haven't written it yet.)

It would have the same "user filter" aspects as a regular Wiki (you'd have to have a user-agent that could do HTTP PUT and understand how to use it well enough to edit the pages) but it would be a lot easier to edit. I wonder what people would do with it.

Stupid

mglazer:

"The greatest purveyor of violence in the world is my own government." - Rev. Dr. Martin Luther King, Jr.
26 Dec 2002 (updated 26 Dec 2002 at 18:55 UTC) »

The trust metric seems to have this problem with persistent inflation. Maybe all certifications should decay? That would enable people who were continuous participants in the community to maintain higher ratings than those who drifted in and out rarely.

The problem that I see with this is that it's difficult to make it from "contribution on external project" to "advogato certification", so you'd end up with the opposite problem (masters who hadn't posted in a month would find their certifications dropped to "apprentice" and have to get people to re-certify them first). This implies that we would need reputation management tools integrated with our source control or email tools.

Maybe not such a bad thing, but I'm not sure where to begin.

Update: Perhaps a fealty system is a way to improve the ratios of rankings here. In order to be a master, one must have a certain number of journeymen working with you; ditto with apprentices. That way a ranking can have some additional values associated with it that gives it strength: he is a master vs. he is my master.

MichaelCrawford: Let me first of all say that I sympathize with your wife's dilemma. Software is in an abominable state these days, and unexpected crashes are a fact of life for far too many people. My significant other has had similar problems with Linux distributions in the past (and in fact, just this afternoon I was helping her debug a graphics driver that was broken by an automatic upgrade).

I know that you've been told a number of times to file bug reports, or to contribute to the Mozilla project, and you seem averse to those suggestions for various reasons. I can also sympathize with your limited resources.

However, your bugzilla and advogato postings belie a certain ignorance about the economic model behind free/open source software. You regard yourself as an informed user, outside the Mozilla organization, offering a bit of helpful advice and concern for the quality of their product.

Here's the magic you're forgetting: no-one is outside the mozilla organization. You are not an independent observer in the "linux" (by which I take it you mean "free/open source software") quality continuum, you are an actor in it. Apparently, you have a goal: to improve the quality and advocate the use of this software. Perhaps you believe you can improve it by agitating about the general concept of improvement on message boards. Based on your Mozilla/IE comparison I take it that one of the milestones you hope that you will reach is to improve the quality of free software to that of its commercial counterparts.

Assuming that your logic is basically sound, I can work backwards from these observations to 2 assumptions that I believe you must hold:

  1. Quality of free software is lower than that of the proprietary alternatives. (Therefore it is very important that it improve soon.)

  2. The quality of free software is this low because the developers are not aware that there is work to do, or they are wrongly prioritizing their work because they do not care about quality. (Therefore it is important that you tell people about it.)

The evidence for your first assumption is poorly presented, as it is anecdotal, but I'll grant that this is at least possible. If this were the only point of disagreement with the mozilla developers, I'm sure that they would have simply said "Well, it works for me, so something else must be wrong with your system.", and you would have easily agreed to disagree.

(Personally I find it hard to believe that IE crashes less for your wife than mozilla. This has not been the case for me for the last year and a half, at least.)

The problem with your second point is a lot worse. The developers of the Mozilla project, much like yourself, are well aware that there are many, many bugs that they need to fix. Your posting (and many like it) is, far from encouraging quality to improve, is impeding any progress in the mozilla project, in three ways.

You give a negative impression of the project's quality to other developers in a completely biased way. Why is mozilla crashing for your wife? Is it the mozilla developers who are incompetent? Have you fallen into the "Power User" trap, modifying all of her system libraries to the point where you've caused the problem yourself? Is there some other software package that's installed of a version that Mozilla was not tested with? You're not helping your wife file systematic bug reports so we don't know. All we know is "mozilla sucks", which is unhelpful to anything but preventing people from wanting to work on mozilla. This further limits the already-limited resources of the mozilla developers to deal with quality problems.

Filing it as a bug report saps the time of existing mozilla developers who could be addressing more serious bug reports. Many of the problems your wife is having could already have been reported, but movement is busy answering your vague report, so he won't be moving on to fix the double-free() which is causing the crash. (If, indeed, the crash is mozilla's fault, which as I said above, we don't know.)

Last, but most certainly not least, it saps the motivation of existing Mozilla developers to do future work, by making them feel that their work is unappreciated. User appreciation is a huge reason for working on open source projects, and it's much rarer to get people who rave about how cool something is than who complain about it. This isn't to say that we can't ever criticize open-source projects, but it is important to balance that criticism and keep it constructive. The open-source culture has long since evolved the etiquette which says that all bug reports should be as technical as possible, to present a developer with a challenge at the same time as a criticism.

I hope that this explanation has been more informative than the blithe "stop submitting vague reports" that you have objected to. So please stop wasting the Mozilla developers' time by promoting your LinuxQuality project before it's done anything useful. (BTW, here's a bug report for you: linuxquality.sunsite.dk doesn't even resolve, from here.) You are not having a "a greater, positive effect on the community as a whole through the kind of quality advocacy role that [you] play". I could have been writing unit tests, but since your article was brought to my attention, I spent an hour trolling on advogato instead.

PS: Mozilla rules, but when are you guys gonna support XHTML in composer, huh? :-)

The inflation of certifications on Advogato disappoints me.

I really wish people would read the certification guidelines before certifying others as "Master" willy-nilly. To be honest, I think that if I deserve a certification as "Master" it's borderline: it hinges upon the fact that Twisted Python is a good open source project, "one that many people depend on, or one that stands out in quality".

I appreciate my peers having given me this confidence, but it is sadly not meaningful. A brief troll through the Advogato network revealed to me maybe 15 or 20 people with "Master" level certification who had made no contribution whatsoever to the open source community, as far as I can tell. (No, a sourceforge page that says "This project has not released any files" does not qualify as a "contribution".)

If you read this, please take some time to lower any certifications you have made that you know are inflated. If you feel you might offend people by ranking them lower than the norm, look at it this way: the overall trust metric may be hopelessly inflated, but your personal certification should still be worth something. It is still quite a compliment, I think, to have people thinking I am a Journeyer! After all, "Journeyers are the people who make free software happen."

The only reason my purple-y banner is of any value to me is because raph ranks me as Master, and it's not because he's close to the root: it's because he's someone who I do not know, and who has very few Master rankings in a very long certification list.

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!