1) Alex Aiken - applying to grad school
2) Richard Stallman - copyright vs. community
so apply everywhere you can - letter writers just send the same thing, you can slightly modify your application letters.
top 4-5 schools - very competitive; element of randomness/personal choice
applicant pools aren't so uniform either - geographic preference, interest in different programs, etc.
so apply to places that aren't the top
be able to answer: why do you want to get a phd/masters? (nothing better to do? like school, and not allowed to be an undergrad anymore?)
if you're interested in the money, stop at a master's degree (it's more years of earning money, and you don't make that much more with a phd)
if you're interested in academics, then you need a phd (academic - prof. or researcher, industrial research like David, highly-technical person at startup like CTO; occasionally positions in government)
if you're uncertain, apply to the phd program - you can always stop at a masters (hard to change from masters to phd, though)
think about different programs you're interested in in grad school.
grad school is a big commitment - 4-6+ years. if you think you can do that, it's good to have an idea of what you like
he says "grad school is a blast" - big research universities often pay all student fees and give you a stipend, and then you can pretty much do what you want (as far as your field goes)
filling out an application -
2- GRE's (big change this year - new analytical test) - take within a month or two of application deadline ... **Anind said general is important; subject can come later ...
if taking the CS GRE is "highly recommended" and not "required" like at Berkeley, keep in mind that the more information you give to the school, the more sure they will be of your
3- letters of recommendation -
- best letters from professors in CS: they know what it takes. most of what admissions committees do is reject people, not accept (200 to accept, 2800 to reject) - they look for the really glowing letters from people they know and trust. if you're working in a research group, let the professor, not a grad student, write the letter!
- other good sources are industrial researchers - they see fewer students, but they're practicing researchers with phds.
- letter from a professor about how you did in a class is okay, not the best
- bad source: letters from industry - research skills are very different from research skills, and caliber of Cal students is usually much higher
- letters from professors outside EECS also not so good: don't understand what it takes in computer science; can be too easily impressed by relatively simple skills.
4- statement of purpose. generally fall into 3 categories:
- 10% show the committee you can't write English ... goes against you (it's a skill you should already have)
- 10% stick out as really cogent - well-written, says something new.
- 80% are in the middle - okay, but don't really stand out. don't count for or against.
don't try to be funny - usually is annoying. if you don't know what you want to do, don't pretend you do - experts are evaluating you, and they can tell you don't know what you're talking about.
apply to lots of places, then you'll hopefully get accepted at a FEW and have choices.
VISIT the places you get accepted to. each school thinks about CS in a different way, and has a different atmosphere.
if you have a specific interest, go to the place that has the best person/people in that field, and have them be your advisor (it doesn't matter where it is)
if you don't have a specific interest, go to a school with lots of strong fields - then when you figure it out, you'll have a good chance of getting a good professor.
statistically, the longer you're in industry after bachelor's degree, the less likely you'll go to grad school
people who do go out for a few years and come back make really good grad students - clearer idea of what they want to do (and they're more sure they want to be there), and clearer idea of how software is really built and really works.
different schools put emphasis on different parts of the application - but every piece is important. at Berkeley, they focus on keeping them as equal as possible; others may focus more in grades.
ask if prof's willing to write a (GOOD) letter ... yes or no
if they say yes, ask what they want (transcripts, resume, etc.)
ask WAY in advance, so the professor will have time to get it in on time!
many students go back and ask for letters of rec. several years after working for them
pass-fail classes - not much difference. (may be better than getting lots of D's ...)
very important point -
committee is looking at evidence of research potential. if you've done research and published a paper or filed a patent, definitely mention it! (get it in the letters of recommendation, statement of purpose) - even if it's just submitted, mention it
graduating early/late doesn't matter
undergrad. teaching experience doesn't really count as a big positive, just a small positive - unless you're an outstanding TA (e.g. can communicate, responsible)
what does "free software" mean? it's not zero-pay software ... it means you have freedom in using it, studying it, changing it, redistributing it. So how does that translate to chairs? have to physically BUILD more chairs, while it's just a matter of copying bits. It is free in other senses - modify and use it as you want. They're free in all ways except copying.
What else does it make sense for? novels? recorded music? what rights should peopl ehave in using their copies of these things? when we think of what "free information" means, we can't reason by analogy - it can suggest solutions, but it invariably breaks down at some point. you have to go through the same thought processes as you've gone through before, in a different context.
tied with questions of law, which decides if we DO have these freedoms
so question of what should be free comes down to copyright.
ancient world - books copied by hand. anyone who could read and write can copy just about as well as anyone else. but 10 copies took 10 times as long as 1 copy - no scale. and making modified copies was as easy as making an original - so there were commentaries, and compendiums. (ancient works have been lost, but fragments survive in compendiums)
gulf between copying and writing was not unbridgable
there existed no concept of "copyright"
printing press - more efficient, but not uniformly - takes time to set type, but not to make more copies once the type is set. books would only be made at certain locations, and shipped around, and others still made copies by hand (rich to have fancy books, poor had time to copy but no money to buy)
copyright developed with printing - Italy, 16th century: ask ruler for a monopoly on printing a certain book; England: censorship, informal agreement among publishers; later, copyright is encouraging authors to write.
US constitution - congress shall have the power to promote science and arts by giving authors exclusive use for a limited time
so copyrights are given not for author's sake, but to modify behavior of possible authors, getting them to write more, for the benefit of society - more conversations about ideas. copyright exists for a public purpose - benefits to authors is just a side-benefit. (and the supreme court has backed this on occasion)
publishers - deny this exists :~)
1900 - printing was more efficient, printed books were cheaper, even the literate poor (cheaper to buy than make - hand-made books disappeared, and people forgot about individual copies
1900 - copyright only covered publication, not copying - industrial regulation for businesses with printing presses, not everyday reader (before the days of xerox ...)
painless - didn't restrict public
easy to enforce - easy to find publishers (they put up ads to sell books)
beneficial - public was giving up freedoms that only publishers could exercise. bargain betwee public and authors - everyone benefits
now - age of computer networks.
digital information technology - easy for everyone to copy and manipulate information. publishers want it to be difficult to do this, to keep control - ever since this age began, publishers
"digital technology creates problems for copyright" - turn it around :~)
state of affairs more like ancient world - anyone who can use the info can copy it, and can do many copies just about as efficiently and effectively as one copy
copyright isn't an industrial regulation - it's a restriction on the public
no longer painless, easy to enforce, or beneficial - stops you from doing useful things, needs to be enforced by everyone, price we pay for benefit is important (not giving up something useless anymore - in the past, we couldn't possibly mass-copy things; now, we can)
now, makes sense for public to keep freedom it found worthless before -
if public could negotiate directly with copyright holders, we could keep some of this freedom.
we could do this if we actually had a democratic gov. - but now it's gov "of the people, by the flunkies, for the corporations"
few will dare say it's morally legitimate to copy things ... like Napster - look how many users, yet they don't stand up for this use.
now, lobbies to take away additional powers - freedom to lend a book, buy a used book, borrow from library, buy anonymously, keep a book for years and read it many times over.
two-step plan to do this:
1- take away freedoms for e-books (since there aren't any e-books anyway)
2- switch to e-books! first to take it up will be the technophiles, who are just awed by the techie side, and then advertisers and marketers who are itching to find a way to make a profit.
hard to stand up for rights ... webcast of a free-sortware talk in non-free format - are your activities helping or hindering?
step that took away e-book freedom - DMCA - passed without a vote in 1998, because it was so uncontroversial (!?) - who would disagree with giving publishers more power?
people don't want to read e-books -
would like to think that it's because of restrictions of freedom, but it's probably more about how hard it is to read on a screen.
but advent of e-paper - no more used book stores, anonymity, lending
digital watermarking is only good if they can tell who copies the book - so you can't buy a book anonymously. And they want more info. about you anyway, for other reasons.
DMCA banned free software to play DVDs ... you can watch your DVDs, but not for free.
consume-but-don't-try-programming act (can't remember the exact name, but these are the initials) - needed to prevent copying
FCC - prohibit digital TV tuners - "tamper-proof" means the customer might change it, and they want to prevent that.
two ways of making it universal - government law, or industry
"trusted computing" - means they want to trust your computer to disobey you. :~) more like "treacherous computing"
make it universal by the "network effect" - certain programs will only run on "trusted" computers
LP records were discontinued because record companies wanted to sell CDs - charge more and pay musicians less.
both companies that make X86 architecture - Intel, AMD - have agreed to "trusted computing" - makes it hard to buy a computer that doesn't have this, hard to do important things (e.g. view files)
what can be done? encryption and signature hardware is useful when it's under your control, so banning that isn't the answer.
one dimension of the extent of copyright - how much power do you vs. copyright holders have?
another dimension - length of time of copyright
want perpetual copyright - get perpetual copyright "on the installment plan" - every 20 years, extend copyright by 20 years, and it works retroactively.
"Sony Bono copyright extension act" - wife replaced him - puppet of church of scientology, which uses copyright to silence critics
Disney - mickey mouse should never go into public domain. Disney did pay money to those that passed this law.
copyright would better benefit public if it was NARROWER and SHORTER
no reason to treat all works alike - people who advocate stricter copyright say that we need uniformities to get anywhere - otherwise everything "grids to a halt". like saying there's a fundamental difference between drunk and sober, though it's a continuum. They say, in this case you need this much copyright power, so let's say you need to buy one for everything! logic of paying the same for everything, whether a car or a carton of milk. here, we're paying the price of freedoms.
distinctions based on USEFULNESS, not MEDIUM - not books, but subjects
3 broad classes - functional, stands for something, aesthetics/entertainment
- functional - use to get something done: software, recipes, reference books and textbooks, etc. - generally not read for fun, but to get to some end. for one, it's been shown to be much better for people to be able to use it flexibly and even republish a modified version. good to have people paying people to make free software, but free software will exist even without these people
free encyclopedias, free dictionaries, free translators; starting to develop free courseware for primary and secondary school in India.
will we be allowed to have free software? maybe not in the future, but free educational materials will continue to be free.
recipes - paradigm for changing and exchanging ideas; software is just following in their footsteps.
so free to comercially publish a modified version - maybe it'd be required to put in credits, but not required to pay or get permission
- stands for something - say what certain groups want, think, etc. (testimonies, scientific papers, opinions, catalogs of goods) - modified versions aren't useful (distorts opinions), but direct copying is okay. also okay to limit to non-commercial ... this is the freedoms that we MUST have. compromise for commercial use and modification of works - then people will get paid like in the time of the printing press. this would make it again an industrial regulation, easy to enforce, perhaps beneficial (?)
so free to copy, noncommercially, verbatim
- art or entertainment - difficult ... is it important to be free to publish modified versions? maybe not novels ... but others it may: Star Wars, the Phantom Edit was made by cutting and rearranging Star Wars. computer games - scenario is like a novel (though people can clearly modify those by playing the game differently). then we have re-writes, like Shakespeare's plot lines in plays - this is illegal now. are his plays rip-offs just because they used stories that were told before? (Disney does it too, but they don't think you should be able to do it TO them ...)
so split: some free to commercially public (games), others free to copy noncommercially, verbatim
automatic conventions for dividing income?
what about length of time? should definitely be shorter than it is now ... and it's only this long because companies that hold copyrights want it. excuse to extend: 75 years isn't long enough to get a benefit ... but they don't have projected benefits for 75 years.
should be 10 years ... but books often go out of print after a few years, and then they're lost into the annals of the publishers. same with music, only worse - CD sales doesn't benefit band. They get "royalties" from CDs, but their publicity expenses are treated as an advance, so the money from the CDs is just paid into this.
why do musicians sign these? musicians have no clout - record companies are the only gateway to success. only if they sell TONS do they actually get any money - they DO get publicity, and revenue from concerts and from selling records at concerts (because they're the retailers)
if everyone could noncommercially copy, then commercial copiers wouldn't be viable, and musicians would be left with their fans that are doing the copying - and there are other ways to get money from fans.
why should we have laws that restrict us so the music industry can decide what we get to listen to?
another option: a small pop-up box, while you listen to a song, that would send $1 to musicians - support musicians tremendously (of course, only the ones that get listened to ... but then it's the public's choice, not the music industry's) - then the musicians will get ALL of the money. Right now, it's hard to send just $1 - once we have a system to do this easily, then it's feasible
book - free software, free society
digital speech project
give money, but also do WORK! protest rallies, informational pamphlets, "hearings" without protesters
both CA senators are supporters of these restrictions - also Pat Shroeder (sp?). everywhere they show up, there should be protesters. you can't think "there's someone else" - there aren't enough "someone else's"
thoughts on the talk:
who was it that was talking about the Linux "old-boy's club", which prides itself on being oh-so-techie ...
companies like MS and AOL are successful because they're incredibly easy to use. right now, it's too difficult for joe shmoe who knows nothing about a computer to get a linux box set up and on a network, because open-source developers code for themselves or for fellow techies. How to make open-source user-friendly, so more people will use it? (And how important is it that the general public gets on the open-source bandwagon?)