About LWN.net

Paul Everitt on the Python Group's move

What does DC gain by hosting the Python team?

The primary reason is world-class talent. These folks have produced seminal work for a sustained time, and we have had a long and productive relationship with them. From that perspective it is a pretty obvious win.

More specifically, we're getting the company ready to compete aggressively in the content management market. One of our strongest differentiators is time-to-market, and Python is core to our ability to deliver on that story.

At the same time, the PythonLabs guys have some extremely unique experience in some problem domains of importance to us. They'll clearly help us shave months off the time needed to get technology into the market, and the value of that is very important to our plans.

Even regarding the core of Python, we'd like to help it take its excellence into increasingly challenging fields. Obviously we're interested in very large-scale application clustering as one example. Python has distinguished itself in the scripting language field as being suited to larger-scale problems, and we'd like to see it take more steps in that direction.

Finally, we've bet the business on Python. We need to help secure its future and increase its success.

Will Zope become part of the Python distribution? Do you expect the PythonLabs people to work on Zope as well?

Personally I don't expect Zope or even important Zope pieces like the object database to become part of the core Python distribution. Though these pieces are applicable outside of Zope, I believe Guido's emphasis is on keeping the Python core pretty tight. But we'll play it by ear.

The PythonLabs guys will spend a portion of their time working on Open Source software that is important to our efforts. Sometimes this will be Zope itself. Sometimes it will be a component of Zope that we'd like to make more broadly useful in Python, like the object database. Other times it will be some core software like setting resource limits on threads.

There might be other initiatives that aren't really software. For instance, we'd like to share some ideas on how to do documentation, so that documenting Python and Zope will have some similarities.

Sometimes the Python community reaches a conclusion, such as distutils, but the decision doesn't catch on very quickly. We can leverage the fact that Zope is a popular Python app to get some momentum behind new Python standards. This is actually an area Guido and I haven't discussed enough yet.

The acquisition of Ajuba Solutions has shaken up the Tcl world somewhat. Do you see any reason to fear a similar occurrence with Python?

I spend most of my time living in the business domain these days, so I'll try to discuss it from that perspective, with the obvious disclaimer that I'm just interpreting from a distance what happened.

Scriptics took on investment with a business model as a tools company. I've always felt that it's hard to generate returns as a tools company, particularly in a double-niche (Open Source and an out-of-the-mainstream scripting language). A different wave came along (XML/B2B), and they followed the dollars to become Ajuba.

At that point the position for Tcl was probably good. Revenue didn't come from Tcl itself, meaning there wasn't so much pressure to get a return from the language. Instead, the language was a dramatic asset for the "soul" of the company (scripting languages are great for XML-oriented architectures).

However, an opportunity came along for the investors to get liquid, and the purchaser had a different architecture established. Perhaps the burn rate was too high and the capital markets too dry for them to pass up the offer.

As you pointed out in your analysis on LWN [October 26, 2000 Weekly Edition -- ed.], the Open Source world has some lines of defense built-in for these kinds of shifts. Additionally, Ajuba seems to have been proactive to launch a new identity for TCL (TCT) in advance. Since Tcl is relatively mature and stable, all should be well for existing Tcl shops. As someone that wants scripting languages in general to take over the world, I hope things work out for Tcl.

Compare the Tcl changes a bit to our situation. We don't sell Python and we never will. We instead make a Python app and sell services based around it. Our strategic success is hitched to the Python bandwagon, but we don't need to monetize Python for our revenue success. We've seen a lot of revenue and capital success this year, so we're acting strategically regarding Python.

At the same time that mainstream interest in Tcl might be on the downswing due to bailout in the forces behind it, I honestly think that Python has a shot at becoming Really Important by this time next year. A lot of ingredients are there. Python is on the upswing, interesting mainstream Python news seems to come out on a weekly basis, Zope is catching hold, and Digital Creations is doing extremely well from a business perspective.

I can say with certainty that this year's Python conference is going to have a fairly cosmic buzz about it, and absolutely everyone should make plans to be there.

Anything else you would like to see in a story about this move?

I'd like to remind people new to the Python world that we at DC pretty deeply get what's going on in the PythonLabs move, and our responsibility to the community. We've been doing this Python thing for a long time, and been deep in the organization of Python for a long time. More importantly, Guido and I now have a strong relationship built over the course of, gosh, going on seven years now. DC appreciates its responsibility and is deeply committed to honoring it.

At the same time, I encourage the Python community to practice "Trust but verify". That is, be vigilant about Python, PythonLabs, and DC. It's not _my_ language, and though Guido is BDFL, his work is really a manifestation of a talented community.