I do not know if this is relevant, but some info FWIW:
As far as I recall (around 3 years ago), most compute nodes on the EGEE grid
( http://en.wikipedia.org/wiki/EGEE , Eddie talked about this in AP2009
- I did not find Hebrew docs )
were "computer-farm" nodes in universities, running SL3 - a clone of
Redhat Enterprise 3, and this was while RHEL5 (and SL5) was already
available.
The python version was ancient. I do not know what the status is today
(you should ask Eddie) - but I'd very surprised if it is higher than 2.5.
(I guess you'd have to write some infrastructure if you want to support
simulations over the Grid, but it sounds like an interesting direction
if you want to extend your audience to certain research areas).
Some other point - I think that the last free version of EPD, before
they went non-free, used Python2.5 (can someone confirm this?)
This might mean that some Windows users might stay with 2.5 until some
alternative free all-in-one python distro is found. This is not a
problem on modern free OS distros, where a reasonable package manager is
available. But in Windows, you typically have to manually install dozens
of stand-alone packages (many of which do not have an up to date binary
egg for Windows on pypi) to get a reasonable scientific environment. EPD
was a really good solution when it was free (it probably still is, for
those who can afford the current license and price).
Amit
On 29/04/2011 00:42, cool-RR wrote:
Good thinking, I didn't think about that. It doesn't hold everywhere,
for example users of NanoHUB <http://nanohub.org>, which is a major
research service, are forced to use Python 2.5, but it does hold in
most places I would guess, so that tilts the scale a little bit
towards ditching 2.5. Thanks!
And hopefully NanoHUB would upgrade to Debian 6 already...
On Thu, Apr 28, 2011 at 5:27 PM, Yishai Beeri <[email protected]
<mailto:[email protected]>> wrote:
My 2 cents:
Is GarlicSim used in production / deployment settings? Or is it
more a tool that runs in a "development" or "research" state of
mind, e.g. is not mission critical, and if it crashes during the
night no-one gets pushed out of bed?
If the former, the people using it will have much less say about
which distro and version they run it on, and hence you should try
to stick with 2.5; if the latter, your users can easily ensure
they have 2.6, and have (almost) no external constraints dictating
an older version.
On Thu, 28 Apr 2011 23:26:03 +0300, cool-RR <[email protected]
<mailto:[email protected]>> wrote:
That's a bummer. But I see that RHEL 6 has Python 2.6, so
people who use
RHEL 6 would be okay. Also fortunately the new Debian comes
with 2.6. (I was
a bit disappointed that they didn't go for 2.7... But they
seem to be a
conservative bunch.)
Do I care? I don't know... I'll think about it a bit. On one
hand I really
want to ditch Python 2.5, it will make me much happier and
solve the bug
that I'm working on (in the `module_tasting` module...), it
will make the
code prettier in various places and will save me from having
to test on
Python 2.5 all the time. On the other hand if it will prevent
some people
from using GarlicSim, that would be a bummer... So I don't
know, I need to
think about it.
On Thu, Apr 28, 2011 at 10:19 PM, guy keren <[email protected]
<mailto:[email protected]>> wrote:
anyone developing software on and supporting RHEL 5.X (and
CentOS 5.X) -
works with python 2.5. the same could be with other non-recent
distributions.
the question is - do you care?
--guy
On Thu, 2011-04-28 at 13:13 +0200, cool-RR wrote:
> Hey everyone,
>
>
> I'm considering dropping Python 2.5 support in
GarlicSim. This would
> solve a compatibility bug I have now with zipimports,
and generally
> make me a happier person.
>
>
> Do you think that people have generally moved off Python
2.5 already?
> Python 2.6 was released two and a half years ago, but
I'm not sure
> whether 2.5 is safe to ignore already. I understand that
GAE still
> runs on 2.5, which is a huge bummer... Though 2.7
support seems to be
> in the "Features on Deck" category in their roadmap, so
I hope they do
> it soon so 2.5 would die already.
>
>
>
> What do you think? Are there any people here who are
forced to work
> with Python 2.5 for some reason?
>
>
>
>
> Ram.
> _______________________________________________
> Python-il mailing list
> [email protected] <mailto:[email protected]>
> http://hamakor.org.il/cgi-bin/mailman/listinfo/python-il
--
Sincerely,
Ram Rachum
_______________________________________________
Python-il mailing list
[email protected]
http://hamakor.org.il/cgi-bin/mailman/listinfo/python-il
_______________________________________________
Python-il mailing list
[email protected]
http://hamakor.org.il/cgi-bin/mailman/listinfo/python-il
