Older blog entries for schoen (starting at number 133)

Lots of changes at Linuxcare, again.

rachel, I also heard it doesn't rain in California, so I also feel betrayed: this isn't supposed to happen, I'm in California!

I wish jimd and dmandala luck in their new jobs. Friday was the last day for each of them at Linuxcare. I wish some other people luck in their new jobs, too -- those caught me more by surprise.

I missed Daniel Burton, Libertarian Not For State Assembly on KQED-FM Thursday evening. It's a pity; I wanted to see how the Republican and Democrat would react to competition from an anarchist. I hope they didn't laugh; as Penrose says, "whatever they should have done, they should not have laughed".

Where can I get bumperstickers made? I think, in retrospect, I want some "Daniel Burton, Libertarian Not For California State Assembly" bumperstickers (not that I have a car). Or even better: "Daniel Burton, Libertarian Against California State Assembly".

I was actually at a Linuxcare dinner (during which I got to talk to mbp on the phone, though I couldn't hear him that well) which turned out to be at a restaurant right around the corner from a Starbucks where a friend works. So that was a nice co-incidence, and I dropped by and rode a bus back to Market Street with her.

I wrote a message about why there are 24 hours in a day, 7 days in a week, and 360 degrees in a circle -- another friend asked. I love that stuff; I think the history of calendars is the coolest thing ever (almost).

I also bought a copy of The Inner Game of Tennis, a controversial book of sports-training psychology, solely because Doug Engelbart praised it extensively in a documentary Brian Harvey showed in CS 61A at Berkeley.

Useless Telco! I still don't have my phone line, because they "weren't informed that [my] roommate already has a telephone line in that apartment" (as they said to crackmonkey when he answered the line in question).

lilo, I still work at Linuxcare. But you can't call me because I don't have any working phone!

Did anybody notice that censorware opponents got one of only two exemptions to the DMCA anticircumvention rules granted by the Librarian of Congress? They may use (but not traffic in) circumvention devices to find out what censorware blocks.

"I hate the DMCA, it makes this song illegal..."

Waldo, I see you noticed that. I disagree with you that it's a good thing from the DMCA. Why? Because the Library of Congress did not, and can not, declare that something like CPHack is legal: all they said is that the DMCA's anticircumvention provisions are not a particular legal theory under which people can be sued for creating programs like that.

There's still trade secret law, unfair competition, breaches of contract (heh!) and even traditional copyright law. Remember that the DMCA wasn't even mentioned in the Cyber Patrol case as providing a cause of action.

We now know that, until 2003, the DMCA won't provide a cause of action in any censorware case like that -- but that doesn't mean that there aren't other causes of action out there.

I had a general comment about the lilo/graydon exchange, and it was directed to stefan, on the subject of "throwing up our hands" and strands of thought in the liberal tradition.

But I think I'm a bit late with that because lilo and graydon are continuing to go back and forth on this. So I'll hold on.

Two really good friends came to visit Friday, so that I got to talk to them for a long time and even walk out to the lovely San Francisco Art Institute with them. SFAI might be a little more lovely when it's not in session, though. :-)

Ouch!

I scanned a couple hundred books. It's nice to have hundreds of books handy for scanning. So I'll release locdump and locextract pretty soon, but I should rewrite locextract in Python for speed. ("Rewrite in Python for speed" is not something you hear every day.)

thom, I notice in that deposit library statute that only UK publishers are affected (which is what I thought). Of course, not every book ever published was published by a UK publisher. :-)

Foreign national library deposits wouldn't be required in order to enforce a foreign copyright under the Berne Convention, so, while these deposit libraries may have almost every book published in the UK, they won't have every book published in the world.

And ISBNs are different for hardcover and paperback editions (I guess deposit libraries probably receive only the hardcovers when both are published -- certainly the Library of Congress, which I believe has to pay for all its acquisitions, seems to prefer to get hardcovers rather than paperbacks). And they're different for US, UK, and Canadian editions, a lot of the time.

So I don't know that I've found a title yet in my collection that wasn't in any national library catalog, but, for ISBNs, it wasn't that hard. Now, if I had a way to cross-reference ISBNs, like a function that could return an "equivalence class" list of other ISBNs that refer to editions containing identical text, that would be nice. That would be really useful for people trying to search foreign libraries!

graydon, how much disk space would konqueror take on a system that lacked KDE libraries and Qt?

srl, how about suing the Attorney General to challenge those laws (a la ACLU v. Reno)? The "crime against nature" one wouldn't be overturned (grumble grumble grumble grumble Bowers v. Hardwick, 478 U.S. 186 (1986) grumble grumble grumble grumble), but I guess you'd do pretty well against blasphemy laws.

(Hmmm, one dissent in Bowers uses "victimless" in a somewhat different way than libertarians do.)

Maybe I shouldn't move back to Massachusetts just yet.

In other news

My right arm really hurts.

Lots of computer professionals don't know binary arithmetic. We must teach the world how all reals can be represented as convergent infinite series whose terms are digits in a base times powers of that base! Or we must behave like Socrates and show them that they already know this! Hooray for place value!

I'm going to get a phone on Thursday, with a new phone number and everything.

Baking soda may help remove spills from carpets.

We had a nice meeting about the Bootable Business Card. I think some very good things will happen with it.

This really is not my week. Chevy's gave me chicken in my nachos again when I ordered vegetarian nachos.

I'm guessing that some people think that "no meat" doesn't imply "no poultry".

I would absolutely love to see a vegetarian certification program in parallel with kashrut/halal certification. Come on, entrepreneurs! And then they could have "this restaurant prepares these dishes in appropriate vegetarian style". Which a kashrut mashciach won't certify, because kashrut doesn't believe that just washing something will purify it.

Maybe I should just learn to cook something other than pasta and rice.

That was just the latest in a series of small disasters which have befallen me all week.

In unrelated news, a friend and I actually thought of a <buzzword>B2B</buzzword> startup that would be profitable (no, really) -- but we don't want to start it or tell other people to start it because we think there are a lot of ethical problems. We should have some reporters write an article "San Francisco Bay Area Twentysomethings Refrain From Founding Internet Company" but just not tell the reporters what our idea actually was. :-)

Actually, we could try to get a business method patent on it and then use the patent in some ethical way. Hmmm. But I'd feel guilty just having a patent at all. The worst thing I've ever done with the USPTO was to apply for a trademark (and then I let the application lapse).

As I look at some of the domains that might describe that idea, it looks like some squatters have registered them; do they actually understand the concept?

... my friend already has his own Internet company anyway, and it's actually profitable, and he's following the "Ben and Jerry's model" with great success. So he's not hurting for Internet startups or anything.

I know that Ralph Nader spoke in Davis on Sunday. It's a long story. Maybe some people had a good time hearing him speak.

I was over in Berkeley again, trying and failing to go to a bookstore. We didn't make it to the Exploratorium.

I promise to send something to raph's bibref.org mailing list soon, like Python or shell code to get bibliographic records from the Library of Congress by ISBN. I'm trying to clean this up a little and do some trivial collection management stuff so that it's convenient to scan books and automatically build a catalog from them.

I could probably write something for my father (a bookdealer) so that he could at least do inventory control with barcodes, although probably 80% of his books don't have ISBNs, much less ISBN barcodes. So he would need to print some kind of bar code labels to affix to books without damaging them. But this could help because he does try to list books through Alibris. I always want to automate my father's business in new ways, but it always turns out to be more difficult than it might first appear.

What I should really do is go visit the national libraries of five or six countries and ask them if they would make bibliographic data from their collections freely available in a standardized format for the benefit of collectors and all sorts of other purposes. But I do have over 50 things on my todo list before that. :-)

Maybe somebody can buy R. R. Bowker (never mind that they're not a public company) and turn them into a non-profit. It should only take a billion dollars or so, right?

Oh, wait... we'd have to make an offer to Reed-Elsevier. Hmmm. How much would Reed-Elsevier cost? :-)

Maybe it's better to start slightly less ambitiously.

thom, books are published at a kind of breakneck pace, so it would be pretty hard for any library to keep up. I once heard that the Library of Congress had every book every published, but I think that claim is extremely out of date (perhaps at least a century out of date), since I have books which aren't in their catalog.

To buy every book published -- assuming one started with an up-to-date collection! -- would take at least a multimillion-dollar annual acquisitions budget. This is of course ignoring many other issues like search and transaction costs, cataloguing and indexing costs, and shelving and archiving and maintenance costs -- not to mention the construction costs needed to build the equivalent of several normal library buildings a year. Right?

I've never heard of a library with more than tens of millions of items in its inventory, but maybe...

Also, even Books in Print doesn't provide a complete list of everything published in a particular year. For one thing, there are still plenty of small publishers far outside the reach of the ISBN system. Like IP addresses, ISBNs now unfortunately cost money. But a book even today need not have an ISBN or bar code or CIP data or anything in order to be a book. (Granted that it will never be on the shelves at Barnes and Noble.) I have a religious tract printed just last year -- bound signatures, title, page numbers, author, copyright notice, table of contents, introduction, a preface, even an edition number. No ISBN, no CIP, no subject headings, no publisher. That's right, it was "privately printed".

I think this is a "book", but I bet the Bodleian doesn't have it, I bet no national libraries have it. What would it cost to track down a copy if the tract went "out of print" before it attracted any notice in any guide to printed works? Is this situation any better for libraries or collectors than publishing in the 1700s?

Hi, agntdrake. /usr/dict/words is now /usr/share/dict/words. I think the FHS requires that.

cmacd, I appreciate your suggestion. But I tried several ISBNs, and it didn't seem that the Canadian National Library had anything that the Library of Congress didn't, at least in my collection, which is overwhelmingly titles printed in the U.S.

brg, I now have a copy of Fear and Trembling. I'll read it soon.

vicious, one can write a Python program (GEGL.py?) which invokes sed to do what you want, without mentioning that forbidden character inside the program. I have a demonstration version, but (darn it!) the regexps in sed are annoying and you'd have to use a number of them. Here's a hint: "[%s%s]" % ((chr(76), chr(108)).

Furthermore, there's a short C program which reads stdin and writes stdout and ensures that any use of that character is in conformance with your instruction. It's a state machine with four states: "none", "G1", "E", and "G2". It ought to be be straightforward to describe the transition conditions. Those who have worked with the techniques for writing programs such as grep may even manage to generate those conditions just as a machine might do it.

It may be a very great heresy, but I have no idea what GEGL is.

hypatia, I certified you and now you have Apprentice. I personally think that running a Nomic game should get you Journeyer at least (but maybe not in free software).

• schoen certified PeterSuber as Master.

Actually, maybe a somewhat modified version of mod_virgule would work well as a forum for on-line Nomic games. Many aspects of voting and rules could be implemented on-line. The interesting thing would be that the code would have to modified in response to certain rule changes. Talk about "code is law"!

I'd like to know if Larry Lessig understands some of Bruce Schneier's points (also daw's points) about what cryptography fundamentally can and can't do. Lessig might be a good person to try to explain to the "outside world" why, for example, copy protection in software can't work.

In other news

I'm still putting off my ambitious ioctl documentation idea because I'm still having RSI problems. I appreciate several readers' suggestions.

I managed to break my phone, so I can't get phone calls, and then both zork and pie went down, so I couldn't get e-mail. It was certainly a very quiet end of the week. (At least the phone has remotely-accessible voicemail and I have a backup MX.)

Also, my toilet overflowed (it's been very unreliable). Since it was empty, it's not as gross as it might have been: but how do you get a bunch of water out of a carpet? The carpet is very absorbent, so paper towels don't take up all the water the way they would on a less absorbent surface.

I think it's actually a situation very like heat flow and heat capacity. (The absorbency of a substance is like its heat capacity, and "wetness" is like temperature; volume of water is like heat. And water will flow only from a more wet material to saturate or wetten a less wet material, when they are placed in contact. Of course, that's ignoring the effect of gravity, because this is a relatively small volume of water on a flat surface.)

I think I'm going to the Exploratorium on Sunday.

I was over in Berkeley for CalLUG on Tuesday; not many people came.

My dad sent me a very remarkable book called The Unquiet Grave, published in London in 1945 under the pseudonym "Palinurus". Not four pages into it, I starting frowning and nodding and laughing uproariously.

My mom found what seems to me to be a legitimate use of JavaScript in a web site! (And zork.net, which is down again at the moment, has the world's only known legitimate use of the <BLINK> tag.)

jimd was teaching on Wednesday again; I'll be teaching part of Thursday and Friday.

brg, I still didn't manage to get a copy of Fear and Trembling. I'll get it very soon.

Palinurus says in The Unquiet Grave that Kierkegaard expressed regret at not having become a police spy, a career for which the latter apparently showed some aptitude. I think that's really funny somehow. Of course, I never imagine that famous philosophers (or writers in general) might have done something totally different.

davidm's new company sounds really cool -- congratulations -- but I share mbp's concern about the rate at which Linuxcare has been losing people. I'm trying to think if any of the technical people who were there when I started are still there. (On reflection, the answer is "yes".)

thom, I took a look at the British Library's OPAC 97 service, but it didn't have anything for either of my first two items that were missing from the Library of Congress collection (ISBNs 0140130470 and 0849304806). So I'm not that encouraged by that; I don't know that their collection is stronger than the LOC's, except in materials published in the UK.

Tragically, Amazon.com has almost all of the ISBNs; I assume they are providing an interface to Books in Print.

It seems that GNU barcode has full details on Code 39 barcodes (which I reverse-engineered once upon a time), and the ability to print them to PostScript. A consequence of this is that I could print out some labels which my CueCat could read, but I don't actually know what good that would do me.

I did make a small sample of a portion of the information I can get from the LOC catalog with my scripts and CueCat. This was generated almost entirely automatically after I scanned the bar codes on the back of 130 books in about twenty minutes yesterday evening.

<Zippy>Yow! I moved to SAN FRANCISCO so that I could write E-MAIL about th' ETYMOLOGY of TUNGSTEN and eat SOY ICE CREAM with CHOPSTICKS while suffering from LOVESICKNESS and TENNIS ELBOW!</Zippy>

jdub, did Raph have an idea for a free on-line bibliographic database?

The Library of Congress stuff is working well -- but they don't have all the books in my collection! (Not just because I have some rare books and foreign books, but because they have a different edition, printing, or binding than I do, resulting in a different ISBN. The Library of Congress doesn't feel compelled to collect every single printing of everything that's printed -- so they are an excellent but not complete source of ISBN data.

What I guess I need is a way to fill in a couple of fields in a form and have a LOC search result with automatic extraction of data in my collection database. That way I can search for books in the above case, as well as things with no bar code and things with no ISBN.

I probably have at most two dozen books that are not in the Library's collection at all. So if I could search on other fields data, thewir data could still help me complete my database faster, and with less typing.

It's obvious a good occasion to write a nice Python class that does arbitrary Library of Congress card catalog searches (or better yet, arbitrary Z39.50 searches) and returns a tuple of BiblioRecord objects. (I don't know what a BiblioRecord object is; I'd have to define that class, too. It might just be a dictionary with particular additional subclass methods for extracting certain "common" fields, like title.)

But I feel a little lazy -- shell scripts are so much fun! Given that I can get the data I want out with sed...

brg, thanks for the suggestion. I'll get a copy next time I'm at City Lights, maybe tonight. I forgot that Hubert Dreyfus was a Berkeley professor; I know him as the author of What Computers [Still] Can't Do, a book which was pretty far ahead of its time in some ways. So both Dreyfus and Searle are at Berkeley. Hmmm.

Waldo, I would really like to try to change the trend of "company sells/gives away interesting device below its own cost, adopts distressing privacy-invasive business plan which is incompatibility with reverse engineering and interoperability; end users make unauthorized use, share tips and get threats of lawsuits". The point is that the companies which are selling these things are not intrinsically bad -- but their business models are very incompatible with certain principles to which a lot of us are fairly attached.

I think a partial solution to this problem is (1) aggressive legal defense for reverse engineering and free software (and I will continue to donate regularly to the EFF or to anybody else who steps up as an aggressive legal defender in that area); (2) trying to show companies that they don't need to have an antagonistic relationship with their customers, or to use intellectual property lawsuits to keep control. It would be nice to show companies that they can still make money even without such, um, antisocial behavior.

The problem is that the traditional way for hardware vendors to make money is by selling products at or above cost. Accordingly, all the companies now providing hardware for free or for a nominal cost have some kind of "gimmick" -- and this gimmick is not very likely to be free software friendly. After all, if you could actually do whatever you wanted with a piece of hardware, how could the vendor make money after you got it?

Well, lots of ways -- so how to find those ways and proclaim them from the proverbial rooftops?

At work, jimd is teaching a class and I'm sitting in. Jim is an official Unix wizard and bearded programmer.

I'll be back in Berkeley tomorrow for CalLUG; I'm also going to try to get rasmus to speak there in November, and some other people at other times. (daw comes to mind; I have a whole list. CalLUG hasn't had a regular speaker series in some years, but I think it would be a good thing.)

Dar Williams says that "life is as hard and as easy as they say"; Vergil notes "et quae sit sententia posco".

Ouch, ouch, ouch, ouch, ouch.

The CueCat is too cool. It worked out of the box with a free userspace driver -- and I managed to scan UPCs, books, and even my old college ID. (It was identified as a "C39" bar code. Who can recommend a good book on bar coding?)

I managed to write a shell script to look up an ISBN in the Library of Congress card catalog, which actually seems a little more complete than Amazon.

I could subscribe to Books in Print, but that costs a lot of money. I was thinking of writing a front-page Advogato entry about why there is no free bibliographic database.

It's very possible that I could get a program done pretty quickly to build a bibliography based on ISBNs (using the LOC catalog); then I could easily (say, in one day) catalog all my books published during my lifetime (well, I'm actually slightly older than ISBN). Then I'd still want a separate catalog of books with no ISBN. (Currently 497 of the 605 books in my on-line catalog have an ISBN; that might be representative of my collection as a whole.)

I'll write some more scripts to aid CueCat (and other bar code reader) automation of book indexing. Then I'll publish them on my web page and send some details to the people who are doing CueCat bibliographic stuff. I'm sure that I'll manage to do this, because it will "scratch an itch" for me.

Bar codes are fun!

I saw The Caucasian Chalk Circle in Berkeley and had many and varied conversations.

In many people's experience, romantic relationships are so powerful that they have the capacity to (inter alia) utterly destroy many wonderful things. I wish I could go back in time and warn some people about that aspect. Maybe they are like nuclear power; I was listening again to "The Great Unknown" by Dar Williams (I wish she'd done that one in concert too). My old .signature file quotation came from there:

They said Look at the light we're giving you
And the darkness that we're saving you from

I seem to have survived that party. I even enjoyed it. (This is not to say that I won't still have problems in the future.)

"I've never been to a party with Seth without a demonstration of the Pulfrich Effect..."

My friend gave me a CueCat (and I promptly took it apart on the bus back to San Francisco -- proprietary ASIC grumble grumble grumble grumble grumble). The argument is very solid that I own the CueCat outright and can do whatever I want with it: he got the CueCat in the mail as a gift from Wired, and then he gave it to me, and neither of us accepted any kind of license or contract.

I'll try to use the CueCat to catalog more of my book collection without hurting my wrists. If Digital Convergence is good for nothing else, they can help save my wrists. (Maybe I should send them a donation of $20 -- to thank them for the CueCat -- if they stop threatening to sue people.)

I'm going back to Berkeley on Sunday to see The Caucasian Chalk Circle; some of us know one of the crew for that play.

Jick, the best source of information on Bay Area Linux and free software groups and events is Rick Moen's Bay Area Linux Events page. Each user group is different, although there is a good deal of overlap in membership.