[CentralOH] Python 3.x

Brian Costlow brian.costlow at gmail.com
Fri Feb 19 15:59:41 CET 2010


Hi Mark,

I highly recommend Dive Into Python 3, available online or as dead trees.

http://diveintopython3.org/

If all else were equal, I'd say go to 3, because if python is to continue to
thrive, that's the future.

But the larger issue is going to be identifying your target market and
whether you can get python 3 running there, not because of tech issues, but
business ones.

If it's a desktop app you're building with a GUI and it's own interpreter
built in, you have some freedom on your choice. Plus, even if you open
source, building say a GTK app is usually harder for the average unix
sysadmin than say, getting a Django web app up and running. So it's more
convenient to just pay for the app.

Server apps, especially web apps, are a little trickier.

I work for a small ISV/VAR. We both resell and create server software. Most
of our customers who are running the Linux version of our stuff are on Red
Hat Enterprise 4 or 5. RHEL 5's latest official python is a 2.4.x version.

You can always roll your own Python and add it to such a system, but:

Large corporate customers (we have some of those) usually won't let you do
that. There's a reason they bought into RHEL 5. Known stable package set
that they are paying support to Red Hat for. Even if you make an isolated
build that doesn't mess with the built in python and it's dependancies at
all, you still usually get a no. Trying to get 'exotic' stuff like Python
and Ruby running on a corporate IT-owned Windows Server is even tougher.
Once again, it's not the tech part that's hardest, but getting past IT best
practice hurdles.

Smaller customers are usually more open (especially if they are buying a
dedicated box just to run your stuff).

In either case, big or small, anything that you add the the Linux distro
vendor doesn't support, you will have to own, support wise.

However (segueing into the money part of your question), the best way to
make a living off of an open source project *is* to support it. Tech
support, installation, training etc. You can also make money developing
features for specific customers where they need something right now. But
it's hard to get paid just for the core product.

Either the Red Hat model (open source the whole thing, sell support) or the
old MySQL model, dual license. But if you dual license, companies paying for
the commercial license will still expect prime support.

But since you are targeting VARs, *they're* going to be the ones making the
support dollars. So now you have to evaluate what you bring to the VARs. How
hard is what you are doing? Most of the commercial apps we resell we
wouldn't develop on, even if they were open. They're complex, require a
larger team than would be cost effective for us, and the developers have a
lot of domain knowledge about the business. Plus we need them for fall-back
support on the really tricky stuff.

If those kind of things apply to your app, you might be able to open it up,
and still sell it wholesale to VARs.


On Fri, Feb 19, 2010 at 8:13 AM, Mark Erbaugh <mark at microenh.com> wrote:

> What are people's favorite references for Python 3.x for experienced 2.x
> users?
>
> I'm contemplating converting a proprietary non-Python system that I've
> obtained the rights to to Python. The end result would be distributed as
> source code (not sure if it would open source or proprietary). The intended
> market is VARs who customize the system for their clients,  although in many
> cases the current system is used as-is.  The current project has been around
> 15 years, and i would hope for a similar run for the converted version.
>  Since I would be starting from scratch, code-wise, does it make more sense
> to start with Python 3.x?
>
> Thanks,
> Mark
>
> *I would like to be able to distribute the system as open source, but it
> even though I was one of the developers of the current system, it will take
> hundreds of hours of my time to convert and I'm looking for something that
> will generate an income stream for me.  Does anyone have any experience or
> suggestions on making a living (or a significant part of one) distributing
> an open source package?
> _______________________________________________
> CentralOH mailing list
> CentralOH at python.org
> http://mail.python.org/mailman/listinfo/centraloh
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/mailman/private/centraloh/attachments/20100219/a95f671b/attachment-0001.htm>


More information about the CentralOH mailing list