Adding Asian codecs to the core
Tamito KAJIYAMA recently announced that he changed the licenses on his Japanese codecs from GPL to a BSD variant. This is great news since this would allow adding the codecs to the Python core which would certainly attract more users to Python in Asia. The codecs are available at: http://pseudo.grad.sccs.chukyo-u.ac.jp/~kajiyama/python/ The codecs are 280kB when compressed as .tar.gz file. Thoughts ? -- Marc-Andre Lemburg CEO eGenix.com Software GmbH ______________________________________________________________________ Company & Consulting: http://www.egenix.com/ Python Software: http://www.lemburg.com/python/
M.-A. Lemburg wrote:
Tamito KAJIYAMA recently announced that he changed the licenses on his Japanese codecs from GPL to a BSD variant. This is great news since this would allow adding the codecs to the Python core which would certainly attract more users to Python in Asia.
The codecs are 280kB when compressed as .tar.gz file.
+0 I like the idea, am uncomfortable with that amount of space. -- --- Aahz (@pobox.com) Hugs and backrubs -- I break Rule 6 <*> http://www.rahul.net/aahz/ Androgynous poly kinky vanilla queer het Pythonista I don't really mind a person having the last whine, but I do mind someone else having the last self-righteous whine.
Aahz Maruch wrote:
M.-A. Lemburg wrote:
Tamito KAJIYAMA recently announced that he changed the licenses on his Japanese codecs from GPL to a BSD variant. This is great news since this would allow adding the codecs to the Python core which would certainly attract more users to Python in Asia.
The codecs are 280kB when compressed as .tar.gz file.
+0
I like the idea, am uncomfortable with that amount of space.
Tamito corrected me about the size (his file includes the .pyc byte code files): the correct size for the sources is 143kB -- almost half of what I initially wrote. If that should still be too much, there are probably some ways to further compress the size of the mapping tables which could be investigated. PS: Tamito is very thrilled about getting his codecs into the core and I am quite certain that he is also prepared to maintain them (I have put him on CC). -- Marc-Andre Lemburg CEO eGenix.com Software GmbH ______________________________________________________________________ Company & Consulting: http://www.egenix.com/ Python Software: http://www.lemburg.com/python/
M.-A. Lemburg wrote:
Aahz Maruch wrote:
M.-A. Lemburg wrote:
Tamito KAJIYAMA recently announced that he changed the licenses on his Japanese codecs from GPL to a BSD variant. This is great news since this would allow adding the codecs to the Python core which would certainly attract more users to Python in Asia.
The codecs are 280kB when compressed as .tar.gz file.
+0
I like the idea, am uncomfortable with that amount of space.
Tamito corrected me about the size (his file includes the .pyc byte code files): the correct size for the sources is 143kB -- almost half of what I initially wrote.
That makes me +0.5, possibly a bit higher. -- --- Aahz (@pobox.com) Hugs and backrubs -- I break Rule 6 <*> http://www.rahul.net/aahz/ Androgynous poly kinky vanilla queer het Pythonista I don't really mind a person having the last whine, but I do mind someone else having the last self-righteous whine.
Aahz Maruch wrote:
....
Tamito corrected me about the size (his file includes the .pyc byte code files): the correct size for the sources is 143kB -- almost half of what I initially wrote.
That makes me +0.5, possibly a bit higher.
We really shouldn't consider the Japanese without Chinese and Korean. And those both seem *larger* than the Japanese. :( What if we add them to CVS and formally maintain them as part of the core but distribute them as a separate download? -- Take a recipe. Leave a recipe. Python Cookbook! http://www.ActiveState.com/pythoncookbook
Paul> What if we add them to CVS and formally maintain them as part of Paul> the core but distribute them as a separate download? That seems to make sense to me. I suspect most Linux distributions (for example) bundle Python into multiple pieces already. My Mandrake system splits the core into (I think) four pieces. It also bundles several other RPMs for PIL, NumPy, Postgres and RPM. Adding another package for a set of codecs doesn't seem like a big deal. Skip
Paul Prescod wrote:
Aahz Maruch wrote:
....
Tamito corrected me about the size (his file includes the .pyc byte code files): the correct size for the sources is 143kB -- almost half of what I initially wrote.
That makes me +0.5, possibly a bit higher.
We really shouldn't consider the Japanese without Chinese and Korean. And those both seem *larger* than the Japanese. :(
Unfortunately, these aren't available under a usable (=non-GPL) license yet.
What if we add them to CVS and formally maintain them as part of the core but distribute them as a separate download?
Good idea. -- Marc-Andre Lemburg CEO eGenix.com Software GmbH ______________________________________________________________________ Company & Consulting: http://www.egenix.com/ Python Software: http://www.lemburg.com/python/
"M.-A. Lemburg" wrote:
...
We really shouldn't consider the Japanese without Chinese and Korean. And those both seem *larger* than the Japanese. :(
Unfortunately, these aren't available under a usable (=non-GPL) license yet.
Frank Chen has agreed to make them available under a Python-style license.
What if we add them to CVS and formally maintain them as part of the core but distribute them as a separate download?
Good idea.
All in favour? -- Take a recipe. Leave a recipe. Python Cookbook! http://www.ActiveState.com/pythoncookbook
What if we add them to CVS and formally maintain them as part of the core but distribute them as a separate download?
Good idea.
All in favour?
+1 -- --- Aahz (@pobox.com) Hugs and backrubs -- I break Rule 6 <*> http://www.rahul.net/aahz/ Androgynous poly kinky vanilla queer het Pythonista I don't really mind a person having the last whine, but I do mind someone else having the last self-righteous whine.
Fredrik Lundh wrote:
Paul wrote:
What if we add them to CVS and formally maintain them as part of the core but distribute them as a separate download?
Good idea.
All in favour?
+0.5
I still think adding them to the core is okay, but that's me.
What would be the threshold for doing so ? Tamito is actively working on reducing the table sizes of the the codecs and after what I have seen you do on these sort of tables I am pretty sure Tamito can turn these tables into shared libs which are smaller than 200k. -- Marc-Andre Lemburg CEO eGenix.com Software GmbH ______________________________________________________________________ Company & Consulting: http://www.egenix.com/ Python Software: http://www.lemburg.com/python/
I still think adding them to the core is okay, but that's me.
What would be the threshold for doing so ?
Tamito is actively working on reducing the table sizes of the the codecs and after what I have seen you do on these sort of tables I am pretty sure Tamito can turn these tables into shared libs which are smaller than 200k.
But isn't this set only one of the many possible Asian codecs? I would have no objection to one 200k module, but if we really wanted to handle "asian codecs" I believe this is only the start. For this reason, I would give a -0 to adding these to the core, and a +1 to adding them to the directory structure proposed by Guido. Mark.
"M.-A. Lemburg" wrote:
...
What would be the threshold for doing so ?
Tamito is actively working on reducing the table sizes of the the codecs and after what I have seen you do on these sort of tables I am pretty sure Tamito can turn these tables into shared libs which are smaller than 200k.
Don't forget Chinese (Taiwan and mainland) and Korean! I guess I don't see the big deal in making them separate downloads. We can use distutils to make them easy to install .exe's for Reference Python and PPM for ActivePython. -- Take a recipe. Leave a recipe. Python Cookbook! http://www.ActiveState.com/pythoncookbook
Paul Prescod wrote:
"M.-A. Lemburg" wrote:
...
What would be the threshold for doing so ?
Tamito is actively working on reducing the table sizes of the the codecs and after what I have seen you do on these sort of tables I am pretty sure Tamito can turn these tables into shared libs which are smaller than 200k.
Don't forget Chinese (Taiwan and mainland) and Korean!
I guess I don't see the big deal in making them separate downloads. We can use distutils to make them easy to install .exe's for Reference Python and PPM for ActivePython.
Ok. BTW, how come www.python.org no longer provides precompiled (contributed) binaries for the various OSes out there ? The FTP server only has these for Python <= 1.5.2. -- Marc-Andre Lemburg CEO eGenix.com Software GmbH ______________________________________________________________________ Company & Consulting: http://www.egenix.com/ Python Software: http://www.lemburg.com/python/
Aahz Maruch wrote:
M.-A. Lemburg wrote:
Aahz Maruch wrote:
M.-A. Lemburg wrote:
Tamito KAJIYAMA recently announced that he changed the licenses on his Japanese codecs from GPL to a BSD variant. This is great news since this would allow adding the codecs to the Python core which would certainly attract more users to Python in Asia.
The codecs are 280kB when compressed as .tar.gz file.
+0
I like the idea, am uncomfortable with that amount of space.
Tamito corrected me about the size (his file includes the .pyc byte code files): the correct size for the sources is 143kB -- almost half of what I initially wrote.
That makes me +0.5, possibly a bit higher.
We will be working on reducing the size of the mapping tables. Can't promise anything, but I believe that Tamito can squeeze them into under 100k using some compression technique (which one is yet to be determined ;). -- Marc-Andre Lemburg CEO eGenix.com Software GmbH ______________________________________________________________________ Company & Consulting: http://www.egenix.com/ Python Software: http://www.lemburg.com/python/
participants (6)
-
aahz@rahul.net
-
Fredrik Lundh
-
M.-A. Lemburg
-
Mark Hammond
-
Paul Prescod
-
Skip Montanaro