Fredrik Lundh wrote: > yndesai wrote: > > > Is it that no compiling facility is hindering the growth of python > > in commercial circuit . . . ?
I can see the point of people who are confused about single file executables for Python programs, who are possibly new to the technology and don't know where to look [1] or which questions to ask, and I can also see the need in various situations for people to make such executables. That said, I don't buy into the kind of hypothetical "ISVs need this and that" pontification seen on places like Planet GNOME, written by people who work at companies like Novell. The "commercial circuit" will use a technology (and, in fact, have been using Python for some time) when they recognise the genuine benefits of the technology, and if the technology doesn't deliver exactly what they had in mind, they'll either put in some effort to shape it to their liking or they'll look elsewhere. If none of this activity has any community benefit, I'd argue that there's only so much the community should be prepared to do to "fix" such commercial objections - if what a business wants is valuable enough, that business should be prepared to pay for it. > (why are you blaming you inability to use Linux installation tools on > Python, btw? most basic Python libraries are only an apt-get away if > you're using a sane Linux distribution.) This is true enough, and by packaging one's programs correctly, Python gets automatically brought into the picture when the user asks to install those programs. It's interesting to consider this in the context of the recent Linux Standard Base discussions on python-dev: LSB potentially mitigates issues with shipping executables across different distributions and would be beneficial to those wanting to deploy Python applications in such a way. I notice, however, that the discussion has taken the peanut gallery position of name-calling and mock outrage at the packaging practices of various distributions [2], presumably whilst advocating Python-only solutions like setuptools - something which really isn't going to work well with any large heterogeneous collection of software packages. Moreover, the distributions have to more urgently deal with various issues not yet sufficiently addressed by the Python core developers, particularly architecture issues [3] and licensing issues [4]. Freezing applications has been a fairly well-understood process for the last ten years, but more cooperation with heterogeneous packaging technologies would be far preferable. After all, distributions are actually responsible for a large amount of Python usage, and it would be far better if people actually tried to work with them to resolve some of the supposedly inflammatory aspects of their packaging practices rather than just shouting bad things at them from a distance [5]. A bit of "not invented here" [6] suppression would also be quite welcome, along with taking the needs of vendors [7] other than Apple Computer Inc. into account. Paul P.S. And while a frank discussion [7] did appear to result in a comprehensive exchange of views between Debian and setuptools developers, I'd like to see a bit more understanding for end-users and people who don't want to ignore their system's package management. Python "plays well with others" is a frequent claim, after all. [1] http://wiki.python.org/moin/DistributionUtilities [2] http://mail.python.org/pipermail/python-dev/2006-November/070032.html [3] http://mail.python.org/pipermail/python-dev/2006-November/070043.html [4] http://mail.python.org/pipermail/python-dev/2006-November/070054.html [5] http://mail.python.org/pipermail/python-dev/2006-November/070055.html [6] http://mail.python.org/pipermail/python-dev/2006-November/070101.html [7] http://mail.python.org/pipermail/distutils-sig/2005-November/005500.html -- http://mail.python.org/mailman/listinfo/python-list
