October, 2007


13
Oct 07

red hat/novell patent Q&A

Because people asked…

Q: will you be doing a Q&A on the Red Hat and Novell patents suits?

A: Are you crazy? :) I’ve worked for both companies and managed a team that may well have written or implemented the code at issue. I’m not touching that in public with a ten foot pole. :)

Q: but! but!

A: Go read Mark Radcliffe’s summary. Better starting place than I’m going to give you. :)

Q: any notes you might want to add?

A: the complaint clearly involves not one but three patents; make sure you take your news with a grain of salt and remove your tinfoil hats before reading. :)

Q: seriously? that’s it?

A: I’m sure I’ll end up chatting with SFLC about the whole thing and may at some point have some comments for GNOME, or at least the board. Unfortunately, given the circumstances, I wouldn’t look to me for much more than that, at least until SFLC as my lawyers tell me I should ;)


11
Oct 07

things I got lectured on in law school today

Today I got a lecture on the virtue of TeX. (Not making this up.) (No prizes for guessing which professor gave that lecture.) (The point was that Word is not good in large part because of lack of choice.)


10
Oct 07

‘retiring.’ Hah.

Good to see that Mark Webbink (my summer boss) hasn’t completely left the fray yet. (He’s also teaching at Duke- I’m jealous; Columbia offers no comparable course, to the best of my knowledge.) I look forward to seeing him tomorrow.

From the press release:

“[D]evelopers perform their best work when sound legal advice is available to them,” said Webbink.

Sums up my career change pretty well.

[Everything below is a followup added later...] 

In comments, someone said that this quote sounded arrogant and asked for explanation. I can’t speak for Mark, of course, but in my own mind, a modern open source developer faces substantial legal questions- patent, copyright, trademark, and DMCA among them. The options are:

  • be nervous, and perform less well.
  • be naive, follow the community’s traditional legal choices, and therefore not be nervous, and perform something like their best work.1
  • get good legal advice, make the right legal choices, and therefore not be nervous, and perform their best work.

Being in the first category is obviously bad. Being in the second category has obviously worked well for most hackers for a long time, but may be getting more dangerous. GNOME wants to fall into the third category, which is why the board retained SFLC and has worked on joining OIN. I personally would like as many open source and free software developers as possible (individuals, startups, and large companies) to fall into the third category.

  1. Note that for some developers, usually corporate, being naive and following their company’s traditional legal choices might be worse than this- it might mean not getting involved at all. []

10
Oct 07

on joe on patents

patents!Some comments on joe’s post on Microsoft’s patents:

they’re unlikely to come after Linux distributors for the well-known “mutually assured destruction” reasons

Which is why they are going after Linux users, most of whom have no patents of their own to retaliate with. Besides the original Fortune article, check out this latest Ballmer quote, where he carefully points out that ‘people that use Red Hat’, rather than Red Hat, owe Microsoft money. (Of course, implicit in that is a threat against all Linux users who haven’t bought immunity from Novell.)

The infighting amongst the community (*cough* with Novell) is counterproductive to the larger goal. We need to advocate for patent system reform…

Well, yes and no; reform is important, but Novell’s alliance with Microsoft actively undermines reform by making the current system seem more reasonable. From now on, on both antitrust and patent fronts, Microsoft gets to point to Novell and say ‘we came to an agreement with Novell, so the system works fine- the problem is that Those Linux People are irrational.’ This is probably the second most damaging part of the Novell-MS alliance (right behind the constant release of free-but-not-really code which MS can later use to threaten users with.)

We need to advocate for patent system reform — like getting involved with the peer review that the USPTO has recently launched

While I do agree that advocating for patent system reform is a good thing, Peer To Patent (P2P) is a mixed bag for open source. If P2P reaches its goals, Microsoft and other large players (including some of the trolls) will have fewer — but still plenty — of patents to use against consumers. And those patents will be issued faster and be harder to challenge in court. On balance, probably still a good thing for the industry, but far from a solution to open source’s problem.

Until then, sadly, we have to play the game and get as many defensive patents as possible…

Completely correct, except that there is no ‘until then’, unless by reform you mean ‘abolition of software patents’. Otherwise, likely everyone will always have to play the game; ‘reform’ by itself will not be sufficient to protect anyone, since trolls with a dozen strong patents will be just as threatening as trolls with hundreds of patents of unknown quality.

… while rebuffing Microsoft’s scare tactics at the same time.

Completely agreed. Novell’s deal compromises the community’s ability to do that, though.

Tangentially, if you want to see what the most advanced patent trolls are thinking, this paper (co-written by a brilliant stanford IP prof and Nathan Myhrvold of Intellectual Ventures, formerly Microsoft) is a really interesting read. It deserves much broader coverage and interest than it has received.

Also tangentially, this post is reaching its one year anniversary and is still the top Google hit for ‘Ubuntu patent policy.’ Compare top hit for ‘Novell patent policy‘ and ‘Red Hat patent policy.’ But hey, they throw good parties. :)

[NP: Radiohead, In Rainbows]

[Picture by Flickr user Ioan Samueli, used under a CC-BY-SA license.]


9
Oct 07

two addendums to the radiohead post

Two quick notes on the radiohead post:

  • I forgot to mention that as part of the purchase process (buried in the terms of service, no opt-out possible!) you give Radiohead the right to email you. Being smart, I doubt they’ll abuse it much, but that is a cost to you which you may want to factor in to your purchase price.
  • NIN has announced they too are going label-free. At this point, a major band going it alone no longer gets to count as surprising or interesting. What is interesting next is how long it will be before we see a band become ‘major’ without ever signing to a label first. (Relatedly, of course, it will be interesting to see what being ‘major’ means in a post-CD age- mega-million dollar tours? mega-million downloads? Something else altogether?)

9
Oct 07

SFLC legal summit

I’ll be there. Drop me a note if you’ll be there and would like to meet up.


7
Oct 07

some free/open services links

I’ve been a little too swamped with school and interviews to do much openservice thinking of late, but it has not been far from my mind. Some links to prove I’m at least reading if not writing:


7
Oct 07

some quick notes on Fedora 8 test 3

Having had a good week, I’m playing with toys a little bit today. (Sadly, not enough free time to make it to GNOME Summit.)1 Some notes, forthwith, from Fedora 8 test 3, with a few supplements as well from a month old Ubuntu Tribe CD.2

  • There is no F8 test3 version in bugzilla. Oops.
  • thinkpad volume keys still broken by default. The most popular line of laptops in the world, guys. And supported out of the box in Ubuntu for a couple of years now. Can’t find this in an open bug, so have filed a new one.3
  • No sign that RH bug 247468 is fixed either, though obviously can’t test that with a liveCD. Every single time I reboot or change the volume these two bugs remind me that in some ways Fedora 7 was a downgrade from Ubuntu c. 2005.
  • still the old, lousy run dialog by default on alt+f2. I would have thought by now that the benefits of binding this to deskbar instead by default would be obvious. (Not Fedora’s fault; more of a GNOME problem. Still sad, especially given that the 2.20 version of the deskbar applet is just as fast as the old dialog.)
  • network manager enabled by default (at least on the livecd.) yay!
  • all tablet functionality still busted by default in Fedora- no rotation and no stylus. Stylus works in Ubuntu’s month-old build, but no rotation there either.
  • new ‘appearance’ capplet is swanky. I do find it amusing that both the Fedora and Ubuntu defaults are apparently ‘custom’. Generally, the new Fedora artwork is great- subtle, professional, very nice.
  • Boot and login feel blazingly fast- I have no idea if this is because I’m reading off a solid-state drive instead of a traditional hard drive, or what, but if it actually holds up, that would be great. Power usage also seems to have improved, according to powertop, though again perhaps because of harddrive usage. Either way, makes me want to upgrade to a SSD ASAP.
  • Once I installed the online-desktop preview, I was impressed. Still some rough edges- most notably that I can’t figure out how to hide it, which sucks on my tiny laptop monitor- but otherwise impressive. Now, we just need someone working on free alternatives for some of the server-side bits.4
  • RH bug 221546 is finally fixed, which is not huge, but nice (especially since it has been broken since F5 or so.)
  • muine works out of the box; it is completely busted in current Ubuntu gutsy. Chalk one up for Fedora on quality/stability.
  • Installing OOo-writer seems to make the liveCD/usb combo barf all over itself- I assume this is something filling up and failing poorly, but I have no idea how to confirm that. So I’ll stick with google docs for my afternoon classwork. Sadly, means I won’t upgrade right now, since I absolutely must have a working OOo and if I can’t test it, I won’t upgrade.
  1. On the plus side, I think I can make it to GUADEC this year. []
  2. Note that fedora makes it very easy to turn a liveCD iso into a USB key (one command line); Ubuntu’s equivalent process is poorly documented and awkward, which is why I haven’t bothered to test a new Ubuntu Tribe as well. []
  3. Could have sworn there was an open bug, but if there was, it has been closed without the problem actually being fixed. []
  4. perhaps someone needs to take a look at installing Joyent Connector at gnome.org? []

2
Oct 07

hypothetical copyright exam question

mystery!

[picture: 'Unsolved Mystery', by ButterflySha, used under a CC-BY license.]

Hypothetical copyright exam question for your late-night pondering: If a poem is derived from statistical analysis of 22 books, but contains no actual direct sequences from those books longer than 2-3 words, is it a derivative work of those 22 books?

Thou shalt be I. . .
Thou shalt be I, and quench the
fire in a pit
dug in the

direction of this rock,
deliberately,
now thy peace, the green weed

and explain the high road from which
miraculously we had
been left unsolved.

from King of Eatable Birds, by Anne Mordeus and the machine.

[hypothetical, though I think probably less interesting, parallel patent exam question: if an evolutionary algorithm creates a better design than humans, and the humans can't necessarily explain why, is the resulting design or process still patentable?]

[Cross-posted from First Movers; comments over there.]


2
Oct 07

two things to know before deciding what to pay for the new Radiohead album

Radiohead is experimenting with allowing you to pay whatever you want for the new Radiohead album, apparently on the theory that you’ll download it anyway, of which they get zero, or buy a CD from a label, of which they get very little.

My friend Iain waiting patiently for a Radiohead concert a long time ago in a place far away. (Image courtesy Nat.)

I thought it might be useful to add two facts for those trying to figure out what to pay for this.

First, what you’ve actually given to Radiohead (as opposed to the distributors) for past albums.

Terry Fisher’s stimulating Promises To Keep has a useful appendix1 discussing what goes where when you buy a CD. Bottom line from Table A.3, p. 262: on average, 12% of the money when you buy a CD goes to the artist; another 4% to the composer. Of course, these numbers are approximate, and averages: some marketing money and other expenses are then taken back from the artists, CDs sold in Canada are even worse (the distributor gets a much larger share), and some megastars get higher numbers in their contracts. But 16% is a good ballpark figure for the percentage of a CD purchase that goes to the composer-musician, which means if you paid $14 for the last few Radiohead albums2 Radiohead actually saw about $2.24 of your $14.

Secondly, the actual costs of the album.

The majority of costs of traditional CD sales go to various overheads- distribution, marketing, etc. So far, Radiohead has done no marketing for this album, except the website. For the digital product, the distribution costs are very, very low. Assuming they use very high quality mp3 files (call it 1/4 gig), and going from the prices charged by Amazon’s S3 service for bandwidth, Radiohead’s cost for distribution will be a little less than five cents.

Combine these two numbers and you get a price floor of about $2.30- if you pay that much for the digital downloads, Radiohead is probably seeing exactly as much money from you for this album as they will have seen from previous albums. Obviously, this is just a ballpark figure, but it is in the right range. Drop below that, and you’ve basically taken advantage of their experiment.

I personally have decided to split the difference with Radiohead; of the $11.70 that is no longer going to marketing, distribution, etc., half will stay in my pocket, and the other half will go to the band (on top of the base $2.30). I’m doing this partially because they were probably hosed by the labels in the past, partially because I want to encourage this kind of experimentation in the music industry3, and partially, frankly, as a bonus for being very good musicians- I like the idea of compensating good artists more than bad artists, instead of having admitted price-fixers set roughly uniform prices across the industry.

But yeah, bottom line: if you give Radiohead three bucks, you can probably safely sleep pretty easy at night that you’re not helping them starve. And if you give more, well, you’ve supported some economically innovative artists. Congratulations.

(And to Berkman folks still reading this blog: yes, this would be a good time for Terry to put out a press release pushing his own thinking in this area ;)

  1. Google Books link. If it doesn’t work, search for “promises to keep appendix: where does the money go” in google books. []
  2. $14 is the average price of their last album from the six stores google lists as selling it []
  3. 16% going to artists means 84% is going to things other than the art itself, which is insane in an age where the marginal cost of distribution approaches zero, and where word of mouth, sometimes computer assisted, like MySpace, is both free and the dominant form of marketing. []

This work by Luis Villa is licensed under a Creative Commons Attribution-ShareAlike 3.0 United States.