Luis VillaRamblings on software, law, and the spaces in between.

responses

Dom: doc is all well and good until we start winning, and then we’ll get sued and FUDed to death. I don’t think it is a coincidence that of all the various formats you can view on a linux box, the best supported are actual open formats like HTML, or that there is the most competition in software in general around things like jpg, html, etc., that are or virtually are open standards.

Glynn: My feedback, actually, is that I’m pleasantly surprised it isn’t 2.8 ;) Congrats to you on getting this out the door- I know this has been a long slog for you and you should be justifiably proud.

Davyd: 24 CPUs now? You’re insane. We love you for it.

other

Read The Big Moo over the past couple of nights while fighting with my dying hard drive (among other things.) As with all business books has some high bullshit factor, but since it is practically a blog in book form (none of the chapters/stories are more than 4-5 pages) if you don’t get anything out of a story, very little of your life was lost. And some fun suggestions/advice/stuff that might be useful.

I’m at a panel on open document formats today. It is a weird mix of brilliance and fluff- Tim Bray in particular is just dead on- he argues for open standards as the precondition for actual, meaningful competition in the tools related to the standard. This is of course obvious on many levels, but I’d never heard it put that way before. Very well said. This particular revolution will soon be webcast by Dan Bricklin, and probably by us too, if we can get the tape from A/V. :)

Tim’s comments basically slammed the current state of the office suite; he didn’t name either one by name but it was clear he didn’t think highly of any of the current options. I hope he’s right and that more widespread adoption of odf means more actual competition on quality and usability instead of just feature creep. Someone argued with Tim that standardizing on one file format reduced choice, but Tim beat him fairly hard upside the head with the state of networking in the ’80s, where there was IBM networking, and Novell networking, and companies I didn’t recognize, and now we’ve got TCP/IP, and while in the 80s people said that standardization in networking reduced choice, it’s clear now that anyone who proposed replacing TCP/IP with a proprietary standard would be laughed out of the room. He feels the same will happen in doc formats in a few years.

David Berlind, who wrote this great article on the state of things in Massachusetts, is here, and is asking some great questions- it is very good to see a journalist who deeply gets the value of open source; better, to see a journalist who knows what questions to ask when someone tries to BS him. We need more journalists like that at all levels, not just technology.

[Tangentially, saw these amusing/depressing posts on openoffice bloat- some hard numbers which I'm sure have some flaws but which aren't out of line with what real people are seeing. Again, OOo needs to learn from the moz->firefox experience- slim down radically and get very good at the very core functionality, and drop everything else, and you'll start to win. Adding a plugin system so that features can be added back if they are actually useful helps too, of course.]

[Bongo looks like it might soon rock my world, speaking of editing documents.]

Met Rob Lucas and Denis Saulnier this morning. They are at Harvard’s Tech in Education program, and Denis is also taking classes at Kennedy School’s Social Entrepreneurship program. We talked quite a bit about playlists, and background and other such. Denis put together a less sophisticated but apparently reasonably successful first cut at lesson plan collaboration based on wiki over at Teacher’s Lounge and is apparently working now on a more sophisticated plan. Should be interesting to see how they develop- hopefully it is something that can be done in partnership with H2O, but we’ll see, of course.

Speaking of new ideas and such at the Berkman- we’re hiring for two Sr. Software Engineers for an as-yet-unannounced project on malware/spyware. The project is very fluid, and anyone who comes on board would have to be both willing and able to start something from scratch. Of course, you’d get the excitement of starting from scratch too, and on a well-funded project that could have a huge impact if it takes off. The Berkman is a great place to work- smart people doing cool things. If you’re interested, apply through Harvard, or drop me a note directly if you have more questions.

Lunch Blogging

Today’s lunch guest is Joshua Schacter, who did delicious; turns out he also did memepool, which was metafilter before it was metafilter. So, wow, he is multiply cool.

The chinese firewall apparently blocks delicious; I think that’s probably a good sign you’re doing something interesting.

There are about 500K unique tags- that growth is leveling off, obviously.

Hard core tech tags are declining, from something like 25% to 17%. Interests of users are starting to get more diverse.

They have bought delicious.com to make the url a little easier :)

Have given some thought to cross-site tagging integration (delicious, flickr, technorati, etc.), but thinks that the semantics for each are different enough to make it more complicated than it looks.

Will probably eventually add ads for non-logged in folks; logged in folks are giving them information and hence are paying their own way.

Describes tagging as the least possible work which still creates some signal.

Later, over dinner: notes that he doesn’t use the word folksonomy, because it implies that it replaces taxonomy. He thinks of tags as words that help you remember, rather than words that help you classify. Memory is the itch you primarily scratch with the tags; discovery (i.e., the result of categorization) is not a primary itch so is a nice side-effect.

Notes that popularity is not the only way that links are interesting- if 9,000 people bookmark /., that’s not interesting, but if 1 other person bookmarks some random site, you and that person have some kind of useful link. So he is thinking about that and how to represent it usefully.

Notes that there are two types of clusters- one in which a human identifies a cluster of things that they believe are similar; and the other in which people apply different terms to similar/same objects.

People search differently than they tag, even though they are sort of flipsides of the same thing. For example, people tag meat products as ‘bacon’, but tag the oracle of bacon as ‘movie’. You’d want a search for ‘bacon’ to return both, which complicates things a bit.

Some interesting links: http://www.siderean.com/facetious/facetious.jsp and http://kevan.org/extispicious. Best: http://blueslugs.com/wordpress/index.php/archives/2005/07/12/tag1-delicious-style-file-tagging/- someone has done tag(1) for file systems. I still want my file selector and save dialogs to be tag-based, dammit… someone get on that :)

Work

For the first time am beginning to feel like my plate is full at work- realized while going through my TODO today that it is getting quite long and quite serious. Ramping up is over and really getting the ass kicked is beginning.

Random Stuff

Apparently Boston will host Wikimania 2006. That should be fun.

Globule is the most interesting new content distribution toy I’ve seen in a while. If it works as advertised, has the potential to make operating mirrors easy for admins and basically transparent for users. That’s a really good thing for anyone who is making ‘amateur’ (read: no funds for bandwidth) content of any sort.

GNOME-y bits

Pretty good article on Novell. Hope none of my friends (in Boston, Provo, or Nuremberg) are personally hit by the cuts that appear to be coming. That said, as a stockholder, I’m happy that the company appears to be focusing itself- hope that once the dust has cleared the right places have been chopped and the right places kept.

Glad to see elijah is blogging again. He has a good link to Gerv on cleaning out old bugs.

Good weekend in New York with the family. Wicked is, I’m sure, an excellent and amusing book, but as a musical, it was… pretty mediocre. Naked Girl on the Appian Way was amusing, if not exactly High Art. Food was good, kids were surprisingly tolerant of MOMA and the Met. Not much more could be asked for.

My direct family will all be out of town for the hurricane, though my step-family on my mom’s side will be home and presumably get wet and wind.

Got my LSAT score. All good. Quite good. Now on to applications.

Have started to get some feedback from folks who saw H2O at educause- glad people liked it enough to play with it. Hopefully the feedback will mostly be positive.

Talk went well. I’ll drop some GNOME jaws by reporting that I basically didn’t talk at all, and Hal did most of the talking :) Think there might be some interesting contacts coming out of it.

Am in an interesting round table right now on open source; on the one hand, lots of people are here, and lots of people are saying very positive things about open and ‘community’ source. And they are being very reasonable about the negatives- they are pragmatic and generally everyone has good, sane responses to the negative stuff. Sadly for me, the one positive that the moderator openly questioned was when I said that use and development of open source matched the mission statement of our institution, and the one negative that was not contested was comments about the religiosity of open source. So we still have issues, both in the PR realm (dealing with the religiosity issue) and in the American economic sphere- getting people to value values is hard.

Looks like someone is finally going to put some resources behind open source textbook creation. Yay. [From Beth Noveck.]

Am at educause today. Some of the presentations are much more interesting than I’d expected- there is hope for education in general; people are doing really cool stuff. They seem to be doing a lot of it in proprietary software (there was an open source talk, but unfortunately I missed it) and often with proprietary content, but at least there seem to be a whole lot of people asking good questions and having good ideas about how multimedia, online interaction, dynamic modeling, MMORPGs, etc. can be used in education. The audience is sort of shocking, though- I’m currently in a student panel, with probably 200-300 people in it, and probably 10 raised their hand when asked if they knew what wikipedia was. They cheered loudly for the one boomer student on the panel (he’s getting a masters on nights/weekends.) It may be that in many cases we have to wait until our generation is in charge to really integrate tech seamlessly into educations, though I definitely give everyone here credit for trying :)

Most random, interesting tech I saw today: clickers are apparently a new toy to get dynamic student interaction in a lecture environment- they can answer multiple-choice questions or rank things 1-10 and the prof can get immediate feedback. Seems like an interesting way of knowing what is going on in your lecture hall and grabbing student’s attention and involvement, though obviously is gameable and has some potentially negative implications for student privacy.

VMware

Philip: no need to worry about my, uh, ‘reviews’; I’ve been very impressed with the vmware beta’s polish. [For those who are at this moment spitting out their milk because Luis was using proprietary software, it was for my LSAT and GMAT review software, and I was very pleased by it.] This is exactly the kind of sophisticated, polished app we need to be showing ISVs as a model for what they can do on our platform. Hopefully we’ll listen to your advice and make it easier for them to do, of course; the interesting response to acrobat was not my post but the responses to my post where people discussed exactly what flaws in our platform would have to be resolved before Adobe could easily remedy the (deep) flaws in their product.

Perhaps tangentially, Philip, now that you’re pimping VIEW, have you and Jody talked about the library of advanced widgets he is always talking about for the GOffice family? Seems like you guys probably have some needs in common…

Travel

Looks like because of the hurricane, instead of meeting my dad and stepmom in New York this weekend, I’ll be meeting my dad, stepmom, and two stepsiblings. Will change the nature of the weekend a bit, I guess, but should still be fun. They’ll also be staying in NYC until at least Monday, because that’ll mean they miss the hurricane altogether.

Orlando, at least the small chunk I’m in for the next 24 hours, is fundamentally not very interesting :)

Plane Reading

Finished The Great Influenza. The author’s writing style is terribly repetitive and hyperbolic, which irritated me pretty quickly, but the material is gripping- the devastating arc of the flu of 1918 (killed substantially more people in a couple months than AIDS has in 25 years, in a world with 1/3rd the population), enmeshed (sort of poorly) with the birth and growth of the American medical research complex. A good read, unless you’re already terrified by our overdueness for another flu epidemic, in which case you’ll lose sleep.

Started on a paper on optimal licensing terms, but then accidentally deleted it mid-read. ARrrrrgh. Apparently evince was perfectly happy to ‘save a copy’ of 0 bytes. The paper had been interesting to that point- it posited and analyzed a model for licensing that was unique (to my knowledge) and which implied the existence of potential types of licenses which we haven’t explored yet. The model basically states (in clarified language) that a platform licensing model has two interesting variables. The first is percentage of the platform X that is free/open. We typically see 0% and 100% as the values for X here, but the game companies have experimented with some other values and are generating some interesting results. The second variable is the amount of time which platform contributors can keep modifications and addons proprietary. Again, the length of time is typically ~0 (GPL) or infinite (BSD, typical proprietary licensing) but apparently the modeling in the part of the paper I didn’t get to suggests that a non-zero, non-infinite value might do a good job of providing incentives for innovation while still increasing investment in the platform over the longrun. To make it concrete, they appear to be positing that if gnome’s source was licensed as ‘GPL, but you can distribute modifications without redistributing the source for a year, but after that year is up you have to redistribute the source as with traditional GPL’ then (for example) Novell and Red Hat might have more incentive to invest than they currently do, since currently, all public development they do runs a decent chance of being shipped first by a distro whose stated goal is not to develop software but only to integrate the development others do.

Started The Right Nation, which is so far excellent.

Finished a rough draft of my trademark paper once I deleted the licensing paper. I’ll try to post it Monday, once I’ve had two more flights to read and review it on, and maybe circulate to a few friends for comments.