[pypy-svn] r3756 - pypy/trunk/doc/funding/negotiations
hpk at codespeak.net
hpk at codespeak.net
Sun Apr 4 22:37:33 CEST 2004
Author: hpk
Date: Sun Apr 4 22:37:32 2004
New Revision: 3756
Modified:
pypy/trunk/doc/funding/negotiations/responses_2nd_april.txt
Log:
revised and checked the responses document
- alltogether shortened paragraphs
- fixed typos
- changed the wording of the "reduce-scope-or-not"
para but left the FIXME there
read the diff ...
Modified: pypy/trunk/doc/funding/negotiations/responses_2nd_april.txt
==============================================================================
--- pypy/trunk/doc/funding/negotiations/responses_2nd_april.txt (original)
+++ pypy/trunk/doc/funding/negotiations/responses_2nd_april.txt Sun Apr 4 22:37:32 2004
@@ -20,36 +20,40 @@
**Action: the involvement and funding mechanism of PBF should be fully described**
-The project coordinator (Alastair Burt, DFKI)sent a mail on the 29th
-of March detailing the future neccessary role of the PBF within the
-PyPy research project. At the same time, three participants, (Holger
-Krekel, Alex Martelli and Christian Tismer) who formerly participated
-via the PBF are now to enter the consortium as independent partners.
-We presume until further feedback that this is a viable model and have
-allocated resources and partners in the DoW accordingly. The PBF is
-to bring in experts on specific problems arising during the process of
-the project. This will decrease the risk of failing. The PBF also
-serves as a valuable vehicle for the dissemination of the project,
-given that it is comprised of SMEs whose business depends on the
-Python programming language. They will be perfectly positioned to
-make immediate use of our new language implementation, the moment it
-becomes commercially advantageous for them to do so, because they will
-be so closely involved in the project as a producer and consumer of
-information. See especially Work Package 14, Documentation and Dissemination.
+The project coordinator (Alastair Burt, DFKI) sent a mail on the 29th of
+March detailing the future role of the PBF within the PyPy research
+project. At the same time, three participants, (Holger Krekel, Alex
+Martelli and Christian Tismer) who formerly would work through the PBF
+are now to enter the consortium as independent partners. We presume
+until further feedback that this is a viable model and have allocated
+resources and partners in the DoW accordingly.
+
+Summarizing, the PBF is to bring in additional experts on specific
+problems arising during the process of the project. This will decrease
+the risk of failing.
+
+The PBF also serves as a valuable vehicle for dissemination to potential
+PyPy integrators and users, given that it is comprised of SMEs whose
+business depends on the Python programming language. They will be
+perfectly positioned to make immediate use of our new language
+implementation, the moment it becomes commercially advantageous for them
+to do so, because they will be so closely involved in the project as a
+producer and consumer of information. See especially Work Package 14,
+Documentation and Dissemination.
**Action: in first instance no workpackages should be removed
as savings should be obtained through economizing the work planning.**
-The consortium has now revisited each work package description and has
+The consortium has revisited each work package description and has
provided a more consistent and more detailed work plan. We
-were able to retain all Workpackages but were more specific about the
+were able to retain all Workpackages but had to be more specific about the
exact deliverables and tasks. This lead to an economized and tightened
work planning.
-However, each of the partners in the consortium now has to accept a
-much higher risk and responsbility for efficiently producing all
-deliverables with the reduced resources. The consortium relies now
-more heavily on interacting early and often with specific communities
+However, due to the budget cuts each of the partners in the consortium
+now had to accept a much higher risk and responsbility for efficiently
+producing all deliverables with the reduced resources. The consortium
+relies heavily on interacting early and often with specific communities
and individuals to effectively mobilise our resources.
XXXFIXME: Do we say we reduced scope or not?? lac+ hpk-
@@ -57,33 +61,24 @@
**Action: provide fully completed and quality checked CPFs.**
We have completed and improved the CPFs. The project office's feedback
-to our consortium-proposal may still make further changes neccessary.
-We also had problems with the CPF editor (import and synchronization
-problems) but hope we fixed all remaining inaccuracies.
-
-We have also discovered a problem which may plague others. In
-Sweden, the correct way to write 'one million, two hundred thirty-seven,
-and five tenths is': 1.237,5
-
-This is precisely the opposite of what is correct in England, or the USA,
-where it is: 1,237.5
-
-Note the swapped commas and periods. Now when Swedes create a CPF,
-using the tool, their locales are configured to produce this X.YYY,Z
-format. And provided they only share data with other people who are
-configured to use this format, all works gloriously well. But as
-soon as they start to share data with somebody who uses the opposite
-nomenclature, because the program supplied either does not store,
-does not check, or does not handle 'what was the locale of the person
-who created this file' then the new user will see 1.237,5 as
-one and 237 thousandths. This explains why you thought we had
-estimated our costs in kilo-euros, when we were certain that we
-had not. It's a locale and language problem.
+to our consortium-proposal may still make further changes neccessary.
+In the course of completing the CPF we noticed a problem which may
+plague others. In Sweden, the correct way to write 'one million, two
+hundred thirty-seven, and five tenths is': 1.237,5. This is precisely
+the inverse of what is correct in England, or the USA, where it is: 1,237.5
+Note the swapped commas and periods.
+
+Now when Swedes create a CPF, using the tool, their country settings
+configured to produce this X.YYY,Z format. If you share this file with
+someone from a different country using a different number format many
+numbers get corrupted. This explains why you thought we had estimated
+our costs in kilo-euros, when we were certain that we had not. It was
+locale and language problem.
**Action: Also the reduction of the language user involvement already
represents a reduction in work and required budget.**
-We still plan to involve language end-users on various levels.
+We still plan to involve potential PyPy users on various levels.
Immediate beneficiaries of the PyPy STREP are predominantly
a) language integrators who want to port and leverage their domain
@@ -98,10 +93,10 @@
- D03.3 will help developers to port existing C-extensions to PyPy
- D04.4 will help researchers to do innovative language research
- D11.1 and D11.2 provide reports and a useful implementation to
- end-users and Axis communications in particular.
+ language users and Axis communications in particular.
- D12.1-3 will provide experimental versions of PyPy to various
- groups of end-users (developers and researchers alike).
-- D13.2 will provide a tool for to support customizing PyPy
+ groups of developers and researchers.
+- D13.2 will provide a tool to support customizing PyPy
to domain specific needs and D13.3 will complement this effort with a
publication.
- all D14 deliverables will target diverse potential users of the new
@@ -110,32 +105,34 @@
game developers both inside (D14.6) and outside (D14.7) the
existing Python community.
-The PBF is a good vehicle for bringing new developers into the project via
-Sprints. This generates a great deal of interest within the existing
-Python development community, interest which is sustained through the
-presentation of papers at various computer conferences, in particular
-PyCon, EuroPython, and OSCON (D14.8). D14.5 will document our experience
-in using Sprints as an integral and essential part of our dissemination
-process.
+The PBF will bring interested parties and developers into the project
+via "Sprints". Those one week meetings serve the dissemination of our
+efforts and are likely
+to raise interest within various communities, interest which is
+sustained through the presentation of papers at various computer
+conferences, in particular PyCon, EuroPython, and OSCON (D14.8). D14.5
+will document our experience in using Sprints as an integral and
+essential part of our dissemination process.
**Action: the body text should be preserved, and a full tracing to
workpackages should be ensured (adding tasks, subtasks, and defined
output). If traceability is not clear, there may come a request to add a
line stating the body text should be considered.**
-We have now fully traced especially the B1 objectives and B6.0 detailed implementation
-chapters in all Workpackage descriptions and deliverables. Task descriptions
-are now more specific and detailed and deliverables have been added according
-to the body text descriptions.
+We have now fully traced the B1 objectives and B6.0 detailed
+implementation chapters in Workpackage descriptions and deliverables.
+In particular, task descriptions of all Work Packages are now much more
+specific and detailed and deliverables have been added according to
+the body text descriptions.
**Action: An explanation on how the envisaged work goes beyond the
state-of-the-art (both in theoretical concepts and solutions and in the
practical implementation should be provided, and how this contributes to
development beyond the Python domain).**
-The appropriate chapter of B1.0 in the TA has been extended accordingly.
-Summarizing we mainly have the following theoretically new approaches
-and solutions:
+The appropriate chapter of B1.0 in the TA has been extended accordingly.
+Summarizing we have the following theoretically new approaches and
+practical solutions:
- Separation of Bytecode interpretation and manipulation of objects,
using a Very-High-Level language to specify semantics
@@ -151,12 +148,12 @@
a way that is inelegant, inflexible, and error-prone.
Each one of these novelties is interesting beyond the scope of Python.
-Moreover, the practical implementation leverages an *industrial-strength*
-language implementation (Python) thus connecting the STREP to a large user
-base. It is possible that PyPy will form the fundation of the "next
-generation" Python implemention commonly refered to as Python 3000. Once
-we have reports and specific versions of PyPy we will reach out to
-non-python communities (see WP14_).
+Moreover, the practical implementation leverages an
+*industrial-strength* language implementation (Python) thus connecting
+the STREP to a large potential user base. It is possible that PyPy will
+form the foundation of the "next generation" Python implementation commonly
+refered to as Python 3000. Once we have reports and specific versions
+of PyPy we will reach out to non-python communities (see WP14_).
**Action: the coordination and management model should be clarified.**
@@ -167,59 +164,53 @@
**Action: Rename the first phase.**
The first phase has been renamed to "Building a Novel Language Research Tool".
-In particular the first phase (D04.4) will provide a Research tool and
+In particular the first phase (D04.4) will provide a novel research tool and
various reseach results on building a highly flexible and modular
interpreter will be published on conferences (see D05.1, D05.3, D05.4).
-FIXME: where do we describe the first phase in more detail so that it
- clear we are new? That should be referenced here.
-
**Action: The DoW will be modified to clarify this [multithreading] issue**.
-The B1.0 objectives chapter has been extended to clarify the issue. The
-WP07_ description, and D7.3 and D7.1 in particular, have been clarified.
+The B1.0 objectives chapter has been extended to clarify the issue and
+in particular the WP07_ description, and D7.3 and D7.1 deliverables.
Summarizing, today there are many complicated issues in current language
implementations such as Java, Perl and Python itself. All such
industrial strength language implementations usually decide early on a
specific threading and memory management model. The PyPy project
pioneers ways to allow making this decision much later and well
encapsulated (D05.3, D05.4). Moreover, massive parallelism with millions
-of threads will be both practical and easy to program without
+of threads will become both practical and easy to program without
consuming operating system resources (D07.1, D07.3, D07.4). This is not
-the case with today's commercially significant languages. This extremely
-significant development is of particular interest to people
-involved in the production of Massively Multi-Player Online Games,
-or complex simulations of any sort.
+the case with today's commercially significant languages. Massive
+parallelism is of particular interest to companies involved in the
+production of Massively Multi-Player Online Games and complex
+simulations of any sort.
**Action: The DoW will be modified to make it clearer what type of tests we are
referring to.**
-The deliverables of WP04 and WP05 now specify more clearly the test selection.
-In general, we select language compliancy tests as the primary measurement
-utility and disregard testing of implementation details specific to the
-current C-Python implementation. The original Python developers will be consulted
-if there is a doubt whether something is a "language feature" or an
-"implementation feature" test.
+The deliverables of WP04 and WP05 now specify choice of appropriate
+tests. In general, we select existing language compliancy tests as the
+primary measurement unit and disregard testing of implementation details
+specific to the current C-Python implementation. The original Python
+developers will be consulted if there is a doubt whether something is a
+"language feature" or an "implementation feature" test.
**Action: The relevant [AOP/logic] sections of the DoW will be extended.**
-Examples are now given in B6.0 to clarify that PyPy aims at researching
-and providing access and interception points to internals greatly improving
+Examples are now mentioned in B6.0 to clarify that PyPy aims at researching
+and providing access and interception points to internals improving
performance and practicality of introducing aspects to the language
-level and enabling Semantic Web applications. The WP09_ and WP10_ have
-been updated accordingly.
-
-XXXFIXME see that this is done.
+level and enabling Semantic Web applications. The WP09_ and WP10_ now
+show more details.
**Action: The DoW will define tasks and deliverables that will be used as vehicles for dissemination to people outside the Python community.**
-WP14_ and WP11_ now define explicit workshops and dissmentiation
-activities. Moreover, several reports and publications have been
-explicitely added to all Work Packages. The B6.0 and B1.0 already
-mentioned various dissemination activities which are now fully traced in
-the DoW tasks and deliverables. Special care has been taken that the
-research report, publication and workshop deliverables will be
-disseminated to non-python communities (see especially D14.1-8, D11.1-2).
+WP14_ and WP11_ now trace workshops and dissmentiation activities
+mentioned earlier in the implementation plan. Moreover, several reports
+and publications have been added to all Work Packages. Special care is
+taken to disseminate research reports, publication and workshop
+deliverables to non-python communities (see especially D14.1-8,
+D11.1-2).
**Action: The DoW will include a justification of the license chosen. Compatibility with the GPL will be addressed.**
@@ -236,17 +227,16 @@
always under an MIT license the project explicitely reserves the right
to *include* GPLed software. In the latter case some versions of PyPy
might have to be released under a GPL license. If PyPy becomes the
-basis of Python 3000, there may be some pressure to release PyPy under
-the Python license. If we develop under the MIT license, then we remain
-compatible to the bsd-like Python license. In contrast, using a GPL
-license would drastically reduce our ability to become the basis of
-Python 3000, since the CPython developers have already rejected the GPL
-as the license for developing Python.
+basis of Python 3000, there will be pressure to release PyPy under the
+Python license which is very similar to the MIT license. In contrast,
+using a GPL license would drastically reduce our ability to become the
+basis of Python 3000, since the CPython developers reject the GPL as the
+license for developing Python.
**Action: The DoW will be modified to avoid this [user] confusion.**
-The complete DoW has been clarified accordingly. The term 'end-user' has been
-dropped in favour of 'language user'. The typical 'language user' of the PyPy
+The term 'end-user' has been dropped in favour of 'language user'. The
+typical 'language user' of the PyPy
STREP are language integrators and researchers. They are also its immediate
beneficiaries. The line between being a language user and a contributor to
the project is designed to be easy-to-cross thus contributing to the effective
More information about the Pypy-commit
mailing list