[pypy-svn] rev 1694 - pypy/trunk/doc/funding
lac at codespeak.net
lac at codespeak.net
Sat Oct 11 01:31:43 CEST 2003
Author: lac
Date: Sat Oct 11 01:31:42 2003
New Revision: 1694
Added:
pypy/trunk/doc/funding/b2.txt
Log:
bed for me.
Added: pypy/trunk/doc/funding/b2.txt
==============================================================================
--- (empty file)
+++ pypy/trunk/doc/funding/b2.txt Sat Oct 11 01:31:42 2003
@@ -0,0 +1,192 @@
+First look at the SWOT table from 'IST Advisory Group: Software technologies,
+embedded systems and distributed systems: A European strategy towards an
+Ambient Intelligent environment'
+
+
+This table describes the situation nicely. However, it falls short of
+having a strategy for addressing the problem presented. We have one.
+
+Our intent is to use the strength in the Open Source Software development
+community (7) and some strong SMEs (8) to combat the Threat of US players
+dominating in development platforms(1). In doing so our principal goal
+will be to overcome Weakness (1) -- Inadequate Structure and Weak Culture
+for transferring and exploiting university research results, fragmentation
+of academic efforts. That's the weakness we are most familiar with, and
+the one we believe we know how to fix.
+
+To put it another way, 'Why is that so much of our research effort
+comes to nothing? Why do the Americans, and not us, take our research
+to market? What exactly is it that goes wrong somewhere between making the
+prototype and making the successful company?'
+
+The answer, we believe, is an inadequate understanding of what it takes
+to become successful. In Europe there is a crucial blindness as to
+what it takes to be successful in the marketplace. Technological
+superiority we have oodles of. It's in the marketplace where we fail.
+And this is in large part because people in Europe still believe that
+good technology sells itself. This is almost never the case.
+
+The Americans know that. This is their big secret. And nobody knows this
+better than the Open Source movement. We routinely create programs that
+are much better than those that are commercially available. And then
+what? Nothing, mostly. This problem in European competitiveness is
+our problem as well.
+
+Over time we have learned, the hard way, that being technologically
+excellent is the easy part of the job. Getting to market is the hard
+part of the job. We believe tthat the European Commission understands
+this problem very well. We do not, however, believe that the EC knows
+how to solve it. Searching over proposals, we found very few Open
+Source ones. This is unsurprising. Perhaps without intending it,
+the EC has made participation by Open Source companies more difficult
+that it needs to.
+
+From our outsider's perspective, we see that the EC thinks that it should
+plough money into some hopeful line of research, and group some large
+industries with some academics, in the hope that the large industries
+will use the research in some of their current projects. There is
+some sincere effort to get SMEs involved, but almost exclusively as
+consumers of research. Small manufacturers, in particular, seem to
+be the 'ideal SME' that the EC envisioned when crafting its plan.
+
+We would like the opportunity to show you an alternative way to attack
+the problem, one that comes from the Open Source and Agile Programming
+world.
+
+First of all, you must start with some SME entrepreneurs and some
+professional educators or communicators. They are essential to your
+eventual success, and they need to be on board even before you begin
+your first technical expedition. This is because the first thing to
+do is to start with a marketing/educating effort. You need to do this
+in order to attract the best people to your project. This is a common
+problem in the Open Source world. Some businesses have embraced
+Open Source, primarily out of greed. They hope to have other people
+develop their software, for free, thus saving them the cost of hiring
+developers. They 'hang up a shingle' -- in other words make a
+web page, and make a few announcements, and then sit back astonished
+when nobody comes. In the Open Source world, it is not the case that
+'if you build it, they will come'. There are companies who do
+succeed at this.
+
+The PyPy development team did things differently. We first decided to
+make a prototype. This has taken us most of the year 2003, because we
+have only been able to do this on our vacations -- about 5 weeks
+total. We did this for 2 reasons. The first is that we were quite
+aware that we were going to ask you for money to enable us to work
+full time creating something useful that had never been done before.
+It would be embarassing and shameful to then fail, because we asked to
+do something that could not be done. A prototype would allow us to
+test our novel concept of Object Spaces. If they did not work, then
+our whole idea was unsound.
+
+Moreover, this would give us the time to greatly publicise our project,
+and attract the best people to it. We happen to be some of the biggest
+fish in the Python-World Pond, but we made sure that we invited others,
+including Guido van Rossum, the biggest fish of all to our development
+Sprints. We worked on our process ensuring that we could work together,
+and that no unexpected personality clashes or conflicts would derail
+the project. We went to conference, after conference, after conference
+and gave papers and informal talks about what we intended to do. We
+made our own website, set up some mailing lists, and discussed things
+not only there, but in the 'python in education' mailing list, the
+'marketing python' (this one is about increasing the market share of
+Python) mailing list, in the Python developer's list, and in the
+newsgroup comp.lang.python. We hung out on our own internet relay
+chat line on irc.freenode.net and answered questions of whoever
+dropped by. We held open Sprints, and let anybody who was interested
+participate.
+
+We made an enormous splash. We are now an extremely high profile
+open source project. We are being watched. If we fail, we will do
+it quite publically.
+
+Why, you might ask, would we go to all that trouble? Because, unlike
+some projects you know, where 'getting the EU funding' is all that
+really matters, and what you do to get it, is secondary -- we find the
+EU funding only a means to our goal. This is because our goal is
+multi-facetted.
+
+First of all there is the straight-forward technical goal. We want to
+make an extremely novel specialising compiler for the Python programming
+language. This goal is technically quite formidable, and requires
+sophisticated expertise.
+
+Second of all, those of us who write software using Python hope that
+our software will itself be improved. This is a commercial interest
+for the PBF, a large stakeholder.
+
+Third of all, there is the marketing goal. We want to make Python
+the most used language on the planet. And that is a much harder goal.
+We had to successfully attract the best people to the project, without
+causing a crisis that would 'fork the project' -- producing 2 hostile
+camps who each point at each other saying 'Mine's the real Python'.
+When we are done making PyPy we want to immediately capture the
+existing user baser of 175,000. <-- should we make an appendix of how
+we calculate that? And then we want to go after those Visual Basic,
+Java, and C++ programmers out there. Python is currently ranked only
+the sixth most popular language on the planet. We would like to become
+number one.
+
+And fourth of all, we want to be the reference project for future
+collaboration between the Open Source and the Agile programming
+communities, and the EU. We think that both groups need each other.
+The EU, in its statement clearly recognises the fact. But it seems,
+to us, unsure how to do this. And there are significant risks which
+the EU seems to want to take with one hand, while want to prevent
+with its other.
+
+For instance, it is not enough to merely have some SMEs involved, in
+some small token way. That will only marginalise them. You will lose
+the very ones you want to attract, those with true entrepreeurial
+spirit, and keep those who are only after a safe and satisfactory life
+at the government trough. If you would like your project adequately
+disseminated, you must involve the SMEs all through the project.
+
+But this is hard to achieve. The smallest SMEs cannot afford to
+dedicate a full person to the project, and will find the paperwork
+crushing. The project cannot afford to rely on a key player who could
+be bought or go bankrupt or become too overworked to contribute at any
+moment of time. What then? There is, of course, a solution. If SMEs
+banded together, and shared resources, they could function like a
+large company. This has worked successfully for Swedish farmers, and
+we believed it could work well for software companies as well. With
+large enough numbers, you can spread the risk around. While at any
+time any PBF member could be in a state of transition, they won't all
+be, and somebody equivalent could be found to take the place of
+somebody who was temporarily unavailable.
+
+This, however, is only possible if great effort is taken at all times to
+keep the process open and transparent. There must be no secrets. You
+cannot get a replacement on a dime unless you are constantly educating
+your larger community. Otherwise, getting up to speed would take too
+long and the project would fail. Aside from the usual web pages,
+conferences, irc channels, and mailing lists, you must make targetted
+FIXME BEA what do you call them? And host open Sprints, keep your
+source open and availabel for download at all times, and use a
+process based FIXME BEA is this correct? Agile development model that
+is robust in the possibility of change.
+
+This is hard. This is very hard. It requires trust, and courage, and
+the sort of Egoless programming which is trained, not read about. It is
+what the EC has claimed it wants, but which it has not fostered.
+
+Thus, from our point of view, we see our relevance to the IST program
+as 2-fold. First of all we are going to give you the AmbiIntelligent
+FIXME_LAURA correct jargon?development platform for networked and
+mobile devices that you believe is essential to European competitivness.
+
+But what we also intend to do is to deliver a model of how you can
+actually get the Open Source community to work with you -- and, of
+course while we deliver the OS community to you, we are also going to
+deliver you to them -- we are going to take special effort to document
+and publicise _you_ to _them_. We are going to teach 'how to crack
+the system and get the EU to pay for you to develop marvels without
+demanding that you lose your soul'. Because, when it comes down to it,
+that is what the OS community wants most. A chance to make a difference,
+change the world, and make it a better place.
+
+If you accept this proposal, and our way of software development, we
+will only be the first in a long line of successful Open Source projects
+that produce the marvels you ask for. And we will stick around and
+teach anybody who comes by exactly how we did it.
+
More information about the Pypy-commit
mailing list