brainstorms

Wednesday May 24, 2006

zphone

i too am playing with zimmermann's zphone...

interesting zrtp draft

Friday April 07, 2006

wily updated

long overdue wily minor update to 0.13.42. tommy pettersson did all the hard work of applying years of collected patches, and other fixes and debugging. [thanks tommy]

Tuesday November 29, 2005

office XML and the art of stage magic

microsoft's office XML moves as stage magic: large props (ECMA, ISO, massachussetts politicians), fanfare, illusion and misdirection.

Many different techniques are used to create misdirection, and all require great amounts of practice to perfect. One technique is the use of natural-looking and confident movements, to disguise any surreptitious manipulations. -- wikipedia [magic (illusion)]

it does not really matter if some in the audience can see through the illusion; their voices will likely be buried under the applause. it is much easier to suspend disbelief than to resist and examine an illusion.

[sun's letter to Massachusetts Executive Office of Administration and Finance.]

Tuesday November 15, 2005

OSS books as necessities...

two new OSS books arrived one after another. i have barely had a chance to skim them. i think karl fogel's producing open source software is very good. it covers [amongst other things] technical, social and financial infrastructure, project communications, release engineering, managing volunteers and choosing licenses. this is a somewhat flat book [important, hard-earned insights are not easy to spot within the text], and i get the strong sense that a more concise 100 page book is struggling to get out. from what i have sampled, i consider it required reading for all OSS starters, including some [cough] old hands. it is also available online, released under a CC attribution-sharealike license.

DiBona, Cooper & Stone (eds) volume open sources 2.0 is a large and mixed bag of essays. [danese's blog entry] i have dipped in here and there [i will admit to reading pamela jones's essay first - it is ok.] and so far find it very informative, but also somewhat flat. [perhaps it is the typeface, or the layout that is giving me that impression. there is a book design issue here. more on this anon]

[are these books printed on acid-free paper? they do not smell like it. sigh]

Wednesday October 05, 2005

style and substance in patent licenses

adobe's tentative dng patent license:

Grant of Rights
Subject to the terms below and solely to permit the reading and writing of image files that comply with the DNG Specification, Adobe hereby grants all individuals and organizations the worldwide, royalty-free, non-transferable, non-exclusive right under all Essential Claims to make, have made, use, sell, import and distribute Compliant Implementations.

compare this to sun's non-assertion covenant for oasis open document format:

Sun irrevocably covenants that, subject solely to the reciprocity requirement described below, it will not seek to enforce any of its enforceable U.S. or foreign patents against any implementation of the Open Document Format for Office Applications (OpenDocument) v1.0 Specification, or of any subsequent version thereof ("OpenDocument Implementation") in which development Sun participates to the point of incurring an obligation, as defined by the rules of OASIS, to grant (or commit to grant) patent licenses or make equivalent non-assertion covenants.

[i was surprised to learn about the patent claims for the dng spec. down that road, i would have much preferred a no-nonsense, sun-style irrevocable covenant for dng.]

Tuesday June 21, 2005

revision control, or the tyranny of adequacy

my good friend david tilbrook once coined the term tyranny of adequacy to describe the continued domination of make and make-like (eg. gmake, jam, cake, ant, ad nauseam) tools for product builds and release engineering. i think the term is now applicable to many other areas of software development, such as revision control. according to the (selection-biased) poll shown at kerneltrap, the most popular [foss] source-code control system appears to be subversion, essentially an adequate next-generation cvs. familiarity breeds no innovation; less familiar but interesting and powerful systems [some, like tom lord's gnu arch, are very idiosyncratic] get fewer users.

i would like to see a poll that covers the entire industry, instead of just the readers of kerneltrap.

reading: david wheeler has a very good overview of the systems shown on this chart.

[note: the graph on the right came out of staroffice. it is adequate for the job, but not more. i am in general not happy with most graphing tools i have found. that will require another blog entry...]

Wednesday June 15, 2005

next morning: groklaw and the art of FUD...

groklaw was quiet yesterday. they must have spent the entire day deciding how to craft just the right article to belittle our efforts while mentioning IBM as well.

Sun Microsystems is the second corporation that needs Open Source. It would like you to help them make some money by writing code for them[*]. You can use the code you write for yourself too, so long as you swear on a stack of Bibles you won't mix it with GPL'd code (the CDDL is incompatible) or help Linux out in any way, shape or form.

i am not the least bit surprised about the tone of the article; groklaw made its analytic modes plain in the past with insults and innuendo. here we have selective Nth reprinting of our sun contributor agreement, and just for added dose of fear, binary-only license and various tone-deaf but "silly me" plain-english translations. funniest part for me is FUD with our joint copyright agreement:

In the case of the Sun Contributor Agreement, my understanding is that you give away your right to control what Sun does with your contribution; however, you are also free to use it any way you like yourself. However, since Sun is the copyright holder of your code, just as you are, it can change the licensing terms at any time it wishes. And, of course, the elephant in the room is Microsoft. Sun and Microsoft are in a patent peace, but you aren't. If Microsoft has patents it could use against you, even if Sun knows about those patents, there is nothing in the agreement or the CDDL that requires Sun to share with you what it knows. It can't be sued, but you can be.

can you follow that neat transition from Sun somehow changing the licensing terms to "and of course" microsoft [why not IBM or any other patent holder? that would be the wrong association] suing the programmer? how we would accept patent-troubled code [there is a clause in the contributor agreement about patents] to our life's work and cannot deal with it, or let the programmer know?

this is all very tiring; best thing to do is to stop taking these barely-clued doses of interpretation and translation from groklaw, and start thinking and asking questions instead.

It can be very dangerous to see things from somebody else's point of view without the proper training.-- Hitchiker's Guide to the Galaxy

[*] "help them make some money" is a link to a straight-forward news item in australianit news. it does not talk about sun making money by having you write code. this is a new twist in blog-journalism: links parading as supporting arguments. if you do not follow the link, you may be tempted to think there is some possibly negative analysis of sun's open source efforts.

note i have disabled further comments on this entry; apologies to those who may have had constructive comments to make. there is now a thread in groklaw about this blognote including my personal likeability (more on that anon); you can contribute your thoughts there if you like, or wait for the next installment. you can also reach me at ozan dot yigit at gmail.

Tuesday June 14, 2005

a big congratulations

a big congratulations to sun and the open source movement. we now have the strongest operating system in the industry open and free. it is a historic, educational and disruptive moment for computing; i am very happy and proud to be a part of one company that has the collective vision to deliver.

music while reflecting on opensolaris:
[classical] angela gheorghiu, the essential angela gheorghiu, london/decca, [jazz] jimmy smith, rockin' the boat, or art blakey & jazz messengers, buhaina's delight blue note rvg edition, [sountrack] ennio morricone, the legendary italian westerns, bmg/rca

funniest entry found in slashdot since the release:

Solaris doesn't stand a chance against *BSD or Linux... their logo sucks! Come on... seriously... what's more cuter than a Penguin or a Daemon? -- jhfry

[evening: ah, finally register has its story and it is even readable. nothing on groklaw.]

Wednesday May 11, 2005

open-source effort is a terrible thing to waste...

there are many acts of stupidity that leave me nearly speechless, and the recent FSF/stallman calls for a fork of OOo (thus avoiding the "java trap") and a rewrite of java are amongst them. this clumsy eWEEK article has the sordid details.

when one is speechless, one's mind is moved into unfamiliar spaces. in this case, i had to think hard about the nature and role of FSF, and about new and creative open source projects in computing, two things at increasing odds with one another.

i no longer think FSF deserves the support and energies of computer scientists and open-source hackers.

in open-source computing, we are not supposed to waste our valuable resources in the increasingly irrational battles against each other and each other's licenses; we are supposed to be professional and smart, and we are supposed to produce good computing solutions to real problems that matter. we are here to make a difference, not slobberingly rework those things that already made a difference.

on the other hand, as my hero kenneth galbraith once put it, if all else fails, immortality can always be assured by spectacular error.

related reading: robert j. sternberg, why smart people can be so stupid
yale university press, 2002.

Monday April 25, 2005

must read: success of open source

i rarely miss important books related to computing, but somehow i managed to not get a copy of steven weber's the success of open source the instant it hit the shelves. how annoying. my friend peter roosen-runge told me about it, and i am now reading. my impression from initial pre-reading: it is the rare kind of book that is smart, well written and deep at the same time. it is a must-read for anyone having anything to do with open-source.

an interesting tidbit i found early on:

In the autumn of 1993, Larry McVoy at Sun Microsystems captured the core of the problem in a memo called "The Sourceware Operating System Proposal" that he prepared for Sun CEO Scott McNealy. The memo began with the proclamation that "Unix is dying" because of duplication of effort around different implementations, leading to high prices; poor compatibility; and worst of all, slower development as each separate Unix vendor had to solve the same kinds of problems independently. [...]

McVoy made radical suggestions for a Unix resurrection. Sun should give away the source code for SunOS 4, its proprietary version of Unix, or simply drop the Sun operating system altogether and adopt Linux instead.

peter wayner's very funny aphorism:

how many open source developers does it take to change a light bulb? 17. 17 to argue about the license; 17 to argue about the brain-deadedness of the light bulb architecture, 17 to argue about a new model that encompasses all models of illumination and makes it simple to replace candles, campfires, pilot lights, and skylights with the same easy-to-extend mechanism, 17 to speculate about the secretive industrial conspiracy that ensures that light bulbs will burn out frequently; 1 to finally change the light bulb, and 16 who decide that this solution is good enough for the time being.

weber's eight general principles of open source process [based on raymond's earlier analysis and his own interviews and observations]

1. make it interesting and make sure it happens
2. scratch an itch
3. minimize how many times you have to reinvent the wheel
4. solve problems thorugh parallel work processes whenever possible
5. leverage the law of large numbers
6. document what you do
7. release early and release often
8. talk a lot

i expect to have a detailed review of this book soon.

open source software is always in beta.

Sunday March 20, 2005

google open source...

google now has open source. pleasantly surprising: it had been very closed until now. [google and open source in one sentence still feels somewhat awkward.]

looked at the sparse hashtable code already; it is very interesting, and more importantly, jolts most programmers' expectations about hash tables; we are really used to simple tables that use chaining. [written in c++ alas. this should probably be re-written in a living language, like java or c.]

i wonder if the google file system code will be opened up.

Thursday February 17, 2005

literacy, osi, etc.

i just read some vacuous, humorous commentary on OSI licenses, but not surprised: these are still the source salad days in some, less imaginative (and incidentally less relevant) parts of the industry. going from not being able to spell G P L without turning blue to commenting on the usefulness of OSI licenses for the future: this is a remarkable development in literacy, sort of like watching my six-year old making his way through a grade-1 reading curriculum:

the source began to glow.
the magic was working.
"oh help!" said biff.

the magic was working.
the trolls got smaller and
smaller and smaller.

related reading, written for adults: lawrence rosen's Open Source Licensing: Software Freedom and Intellectual Property Law. it has excellent coverage of reciprocal (GPL), academic (BSD, MIT etc) and alternative (MPL, CPL, OSL etc) licenses, alternative models to open source etc. highly recommended.

[addendum: i was very pleased to note that russ nelson will be the new president of OSI. congratulations, russ.]

(2005-02-17 20:53:13.0) Permalink Comments [0]

Thursday January 06, 2005

performance comparison: NetBSD vs FreeBSD

gregory mcgarry did a performance comparison between NetBSD 2.0 and FreeBSD 5.3. the paper is here. the benchmarks test core os functionality, scalability and thread implementation. very impressive: looks like netbsd is going far and fast. [i actively use both, but netbsd has been running my core mail/web server and in my voyagers and aging vaio for a long time...]

(2005-01-05 21:22:47.0) Permalink Comments [2]