Date: Tue, 09 May 2000 23:23:11 -0400
| From: | Paul Fernhout |
|
pdfernhout@kurtz-fernhout.com Reply-To: unrev-II@egroups.com Organization: Kurtz-Fernhout Software |
| To: | unrev-II@egroups.com |
| Subject: | Species survival & open source funding etc. |
Eric -
My wife and I went through something like this a few years ago when deciding to finish our garden simulator and then release it as open source. That was six person-years of effort and sacrifice. Except for some advertising / resume value, we never received a dime for any of that work (in fact, it still costs us money each month to make it available.)
At the time, we were especially disheartened by a rejection letter from the NSF for a grant pre-proposal saying basically "just sell it".
Obviously part of the problem was our grant "sales" pitch and communication skills, but another part of the problem is the lagging pace of community awareness of new ideas -- especially among conservative PhD academics who decide about most grant making.
Five to ten years from now funding open source projects by the NSF or such will be common place -- and then all the sharp marketers out there (rarely the Dougs of the world) will get most of the the funds to do stupid things that sound good or reinvent things the Dougs of the world suffered to create years earlier. (The book, "The Seven Laws of Money" by Michael Phillips has a chapter on why most grants go astray.) The hope is in the 5% of funding that accidentally falls into the right hands (against the best efforts of institutional conservatism to prevent it or ignore it or not understand it) like when J.R. Licklider gave money to Doug in the 1960s.
The reason for this is that almost all the useful projects help humanity, but it is a strong part of human nature to help yourself or your local group -- and often those two things are at odds (even when you work for a government or the UN or World Bank). The one that usually wins is helping yourself. The creation of large immortal corporations helping themselves only makes this funding competition more difficult. For example, a government grant could go to fund an idealistic open source programmer for a year, or instead produce a 5% profit for a company making proprietary software. Guess who is in a better position to "buy" the grant with smoozing and lobbying and displaying a "track record" and "qualified professionals" by having PhDs on staff. This isn't helped by the push nowadays for every grantee to become self-funding, so if you can't prove you will turn the grant into a perpetual revenue stream, you probably won't get it (and this in software means making proprietary stuff).
Also, the "peer review" process only works to fund typical work done in a way everyone accepts. Anything really new will fail the peer review process. The originator will starve and move onto other things, and years later the idea is "in the air" and typically a reviewer unconsciously remembers the idea and gets funded for proposing it. The work is then carried out without the full wisdom or context of the originator, and so it is weakened and likely fails to produce a wholistically integrated result. I've heard of this happening many times -- and I do not mean to imply any deceit or ill will on any of the reviewing parties. This is perhaps the nature of social consensus building -- but it is made worse by the increasingly narrow specialization of academics and their "departments" / compartments, which everyone abhors but no one seems to be able to change.
Sometimes you just got to do what you think is right -- like we did deciding to finish our garden simulator in the hopes some people somewhere would learn how to grow their own food in a more sustainable way from it (or at least, that others might make it into such). I've spent the last couple of years paying for it in interest & principal though! That's only money, but it's probably also delayed or prevented us having children. I just shockingly realized on Sunday that (no joking) in a way I have already [in a "pact with the devil" sort of way] given my first (never-to-be)born child to open source type projects. :-( But there are plenty of people who have given more than that for causes they believed in -- be it "racial equality" or "democracy" or "disarmament" or "ecology".
I think of the choice as deciding among "the many scales of self". You've got to decide at each point in time what scale of self (body, family, organization, ecosystem, nation, Gaia, universe) you are going to be "selfish" for. Sometimes that means sacrificing the levels below. (Note this is made more complex by multiple overlapping entities on the same level -- like belonging to a church and a company and a lodge -- and sometimes having to choose between them).
The world is full of people who volunteer -- teaching illiterate adults to read, giving blood, or raising foster children. It's also full of people giving money to flood victims or funding foundations. There is enough slack in there for good things to happen. It may not be happening with Bootstrap as quickly as one might like, but it is happening somewhere.
Any of the digital library efforts are heading towards such things. I especially like the CANIS stuff. "CANIS is developing and deploying unique analysis environments for large-scale information retrieval applications based on discipline and community scale collections."
If I was going to work for someone and was relocatable, I'd strongly considering joining that group [I can't easily relocate because of elderly parents]. (One important concept the CANIS group looks at is how the explicit linking concept is sort of obsolete -- in reality dynamic searches will replace links...)
The good news is if you are good enough to make a major open source contribution, you are good enough to earn enough working half your time (in many month clumps) as a contractor to buy your own time (in other many month clumps). It would be nice if there was a better alternative, but the other possibility is competing for grants -- and good luck if you don't have the right PhD and the right connections (and lots of supported time to write the grants). The other possibility is to find the sharp marketers with the right PhDs and hot connections and work for them for half to a third (or less) of what you might get in industry. But that is still better than the lot many people on this planet have in life. Marvin Minsky bemoans the fact he can't get good graduate students because anyone good can earn $100K+ in industry vs $20K as a grad student. (The pyramid scheme of PhD granting academia coming to a brutal end is perhaps another reason in my opinion. :-) On PhD career issues and career half-lives:
This is a not-quite-fair assessment of Bootstrap/Augment and its funding history but basically as it seems to me based on what I have read: Doug had a great vision similar to a few other visionaries of his time (specifically, Vannevar Bush [Memex - 1950s]
...and J.R. Licklider and even Theodore Sturgeon in the story "The Skills of Xanadu" and also Ted Nelson
Doug got funding from Licklider when Lick was at Darpa and through the help of several bright technical people realized a prototype of that vision -- something beyond anything of the time, and perhaps in a way still beyond us now in some ways. (Read the licklider article I posted the URL for
...to see how Lick funding also created pretty much all the other CS greats of the 1970s & 1980s). His was the 5% of funding (see above) that got misdirected to do something useful -- first trickled to Lick, and then to Doug. Then that funding was cut off as Lick left Darpa and because Doug is a visionary more than a marketer things floundered. What followed was a typical short-sighted business scenario at Boeing. The only surprise there is that Augment got as far as it did.
Doug got recognized decades later after many hard and wearying struggles by a $500,000 Lemelson-MIT prize in 1997
...but presumably the money ($250K after taxes?) probably (I'm guessing) went for either previous expenses or for a few years of Bootstrap operations, or for a trust/investment to support him personally (since income on that would be say $25K per year -- I heard Stallman did this last with his $500,000 Macarthur genius award). I wouldn't begrudge Doug that last choice, and he deserves far better.
Obviously it does seem he could have bought some programmer time with that prize -- and perhaps he did -- but good/experienced programmers are expensive (it's typically $200K+ per year [includes overhead] to have a good programmer in a large organization) and a year's worth of programming from a random programmer might not have been the best investment of the funds. Programmer productivity can easily vary by an order of magnitude and more. So it made sense for Doug to ensure the tiny amount of money went towards broadcasting his message of the hope of Bootstrapping. Unfortunately, getting such a prize probably in a way undercuts his ability to gain support in some ways (i.e. people probably asks why doesn't he just pay for the implementation?). He also got some DARPA money which I guess went into a study.
Basically Doug is focusing much of his effort and the colloquium on making people aware of global issues and the need for bootstrapping, which is good, but alas the software aspect got lost somewhat. [For example, sorry, I kind of lost interest in the webcasts around #5] I do agree with Doug's points about coevolution of organizations and their tools, but the tools is the only real leverage point we realistically have, so that is where I think most of the effort should go. I also think (sorry Doug) the concept of endless levels of Bootstrapping is unrealistic given the S-curve nature of most things and more support might be forthcoming if there was a specific thing or level being bootstrapped towards.
And it seems to me many people on the list and colloquium have participated from an organizational and issue perspectives, and not a coding perspective -- i.e. there is very little coding released.
I like this list as a forum for such (like your own post just now). I feel that personally I have gotten a lot out of participating in it. But sadly, the software part has floundered. Perhaps this is because the spec if essentially a waterfall model -- "code it and they will come" which daunts a free spirited open source developer. In reality almost all good software design is iterative -- exposing unknown requirements in the act of iterations, as in "code it again and some more will come".
And of course, when you don't have much time, there is always the "do" vs. "discuss" choice.
Jim Spohrer had encouraged me to participate, for among other reasons:
In the end I decided to participate based in part on that. I included the letter I sent Jim first at the end of this note, showing how there are similar such projects already afoot in various ways. [Even then, the lack of an open source license or released code worried me...] I still think it is true that Doug brings visibility to a project, and he also brings history (which has both good and bad sides).
I think part of the issue is that Doug from what I've read (not having met him personally) is not and has never claimed to be an aggressive marketer or pushy-type leader (in a human-leader sort of sense). Yet, we all look up to him to provide leadership (fundraising, organization, license commitment) while he probably looks at us to provide the same!
Unfortunately, the realities of fund raising and legal commitments require a meeting of minds and some official decisions with legally binding signatures or licenses. This means inertia. Throw the issue of funding and decisions about becoming proprietary into this mix, and it makes it hard to get critical mass for "open source". Ad Stanford's likely desire to make a profit selling courseware based on the Colloquium and things get really complicated.
The reality of most open source developments is they start with a developer (or a few) with some source the make available under specific licensing terms.
Usually the initial developers then shepherd it for some time (years). That forms the seed from which the effort is bootstrapped. It attracts people who use it. If the tool attracts enough programmers it grows (easier if it is a programming tool, less so if it is not.) We don't have that seed. The Augment code is the most obvious code and it is not available for various reasons. Even if one had a seed to contribute, the "permission to use" indemnification clause and the agreement's one-sided nature would make it legally foolish for one to contribute it.
At this point, the license issue (esp. indemnification) has stopped me from participating further (in terms of contributing code). Actually, I was surprised no one commented on my recent post on it. I plan to continue open source releases on my own related to knowledge repositories along the lines of stuff I have been working on similar to William Kent's model in Data&Reality -- but not as part of the "extended activities of the colloquium" though -- until such license issues are resolved to my satisfaction. Just settling on X11 / BSD revised might probably be enough, and moving code development out from under the "permission to use" indemnifying license (most open source software comes with NO WARRANTY, yet "the permission to use" license we agree to by participating means we provide a warranty to Stanford and BI (and their licensees) for our code contributions).
I write this (somewhat negative towards the current state of affairs) in hopes that things will perk up. In a way the list is in crisis -- months of participation and no code. This is not good... Something needs to change -- perhaps starting with the permission to use indemnification issue... However, being on the East Coast I haven't attended any of the recent weekly meetings, so I can only hope there is a coding ferment there not entirely reflected on the list yet.
In any case, Eric, I know where you are coming from. It is disheartening to know one has so much to give and cultural circumstances preoccupy one so as to force one to give it to a "lesser" cause to ensure one's individual survival. But almost everything worthwhile requires patience, persistance, hard work, and (unfortunately) sacrifice. Doug is a shining example of "character" -- and he has suffered for it (and yet, apparently, not complained too much.)
Too bad we are in a profession where the ergonomics are so terrible a full day at the office injures a programmer enough they have less reserve for doing the same elsewhere (as opposed to say 40 hours a wek of walking around and talking to people -- where you can easily do another 40 hours a week of the same.)
Sincerely,
Kurtz-Fernhout Software
Paul Fernhout
pdfernhout@kurtz-fernhout.com
Date: Wed, 24 Nov 1999 10:13:38 -0500
| From: | Paul Fernhout |
|
pdfernhout@kurtz-fernhout.com Organization: Kurtz-Fernhout Software |
Jim -
I really appreciate your email on this. The Boot strap institue sounds like a great project, and I would love to work with them.
Certainly good points! And don't get me wrong, I think the Bootstrap institute is one of the best projects going...
Some context though:
I was interested in BSCW (Basic Support for Cooperative Work)for a while:
...which is sort of the same thing (I think) as OHS in Python. I got an account on their system a couple years back to try it out. However, they've moved into the for-profit licensing, and so I lost interest in using/improving their code. using/improving their code. http://bscw.gmd.de/Copyright.html (Also, their system seems overly complex for first time users.)
Of course, I'm sure Bootstrap is aware of:
(CANIS == Community Architecture for Network Information Systems) which is a heavily funded effort to support collaborative computing (although not open source). I know someone who is on that project (Les Tyrell, also a Squeaker) and he sure is having fun. I'd try to work there except I'm tied down to the NY area for the moment.
And of course, to a lesser extent:
But they are not open source for their toolbase either (although the results are available in a massive archive of educational material -- I'm not clear on the license for that though).
And an upcoming effort from an owner of VA Linux (though not collaborative yet):
This is a really cool open source project for a clinical record and recall system for community based health care in the developing world:
(since medical records and related decision making are another example of collaborative computing.)
So, why pick Bootstrap to work with?
After more perusal of their site, I see "Open Source OHS Project":
However, I haven't yet come across a specific license, or a related content license. But, there's a lot I could contribute on the presumption it will be some version of open source.
Of course, I think XML is problematic (how do they handle versions of DTDs?) and I'm still not sold on Java. I'm encouraged they had a client application a few years back in VisualWorks Smalltalk (a system I know well and love). However, if they develop clear APIs and well documented storage techniques, language should not be as much of an issue.
As to your points: working with one of the greatest minds behind the computer revolution (on par with or beyond Vannevar Bush)
...is certainly an exciting prospect. And I think Doug is continuing to head in the right direction. (I started calling a version of the system I'm working on "Alexandria".) And I'd certainly like to interact more with you.
As for goals, my model of wealth and security in today's interrelated world [taken to an extreme :-) ] is that I won't feel wealthy or secure until everyone around me feels the same way. Since I think gated communities [or countries] are ultimately ineffectual and rude [see Jane Jacobs work on cities for what makes a safe community], that means every person on the planet needs to feel wealthy and secure. That is what I would like to work towards. At this point, I think a good hope for world prosperity is through making core knowledge freely available, along with a somewhat self-replicating technical infrastructure with which to apply that knowledge. See question 11 of this Final Exam to understand why I think the two (tools and knowledge) are intertwined:
Anyway, I like where the Bootstrap institute is going. If they want me, I'd be happy to given them a chance working full-time mostly telecomuting spending DARPA money on useful stuff that provides a real defense against "want and ignorance", based purely on their web site, track record, and stated goals (and "open source" statement). Otherwise, I'll see what I can do as I have time to contribute to their OHS project part time on my own.
And I'd love to be conference in on one of those phone calls -- (and I should be in my home office on Friday, or let me know when and I'll make time.)
Feel free to forward this note to Bootstrap people. I can send a resume too if they want. Twenty years programming experience, mumble, GPL garden simulator, mumble, IBM Research, mumble mumble... :-)
Kurtz-Fernhout Software
Paul Fernhout
pdfernhout@kurtz-fernhout.com