Blue Sky On Mars

Ben Bangert talks about Python Web Framework Niches, specifically detailing TurboGears and Django. I think he’s on the mark that each framework defines its niche based on what it’s asked to do. The best APIs are ones that solve real problems for applications, and not ones that are just created because something sounds neat or potentially useful.

The part that Ben leaves out is the matter of taste (which I touch upon in a comment). I’ve done a lot of work with Perl, and I still think its a good language for quick one-off scripts. Perl certainly can be used to create large systems, and I’ve worked on large systems written in Perl. Some people swear by it, but I don’t like large Perl programs. That’s purely a matter of taste, and I’m sure this extends to the APIs we chose as much as it does to the programming languages we choose.

Python

TurboGears 0.8a1 is available now!

What’s New

This is a brief summary. The complete information about what’s
new can be found here: http://www.turbogears.org/about/changelog.html

* API improvements based on feedback and patches from the first public release. Seven people contributed patches to TurboGears directly, and there’s been quite a bit going on in the other projects.
* Easier production of XML output from controller methods
* Static file directories are set up in new quickstarted projects
* Updates to all of the main included projects
* IPython is used as the shell, if IPython is available
* Bonjour support on the Mac
* New getting started guide, and command line tool and configuration references and a new site template by Sebastian Jansson.
* Several bug fixes

Python, TurboGears

After the slashdotting, Ian Landsman wondered what the extended effect of the slashdotting would be. Certainly, after serving more than 13,000 views of the screencast on Monday, things have tapered off. The 20 Minute Wiki is still getting 1,000 views a day! (That’s 80GB per day, for those watching their waistline bandwidth usage). I haven’t had a chance to do a total tally, but I’m guessing that the screencast is up to 25,000 views.

The main TurboGears discussion Google Group has added about 100 people since Monday, crossing 350 earlier today. The announcements list crossed 100 today. Mailing list traffic is increasing, and people are coming up with some cool stuff. Interesting things are taking shape for 0.9!

Thanks to all of the people putting their energy into TurboGears!

Python, TurboGears

Every time a prospective customer needs to make a choice, you open up the possibility that the choice will be to head somewhere else.

A couple months ago, Apple offered: the iPod Shuffle, iPod mini, iPod and iPod Photo. I’m not going to count the U2 iPod, because that’s more of a niche gimmick sort of thing. Since then, Apple has introduced two new models, but reduced the number of choices: iPod Shuffle, iPod nano and iPod. These models come in at nicely spaced price points, and it should be pretty clear to a potential customer which one they should buy.

Contrast that with Creative’s lineup. Desparate to win some market share, these guys seem to want to throw a ton of products at the wall to see which stick. At first I thought they had 9 models, but then I noticed the little arrows. They 9 Zen models, 11 Muvo models and something called the “FX200″.

From some reviews I’ve read, some of these models are actually quite good. Of course, it’s impossible for me to remember which of the 21 models those are. Do stores really have shelf space for 21 models, some of which come in 10 colors? How can they profitably produce so many models?

Apple appears to understand market segmentation. They offer an entry level device, an ultra portable skip-proof device and a full-size model with good storage capacity. It’s crystal clear how Apple has segmented their market, which means the choice of product will be crystal clear to the people who are prospective Apple customers. Apple’s 75% share of the market proves that their segmentation isn’t leaving many users out.

I challenge someone to explain Creative’s market segmentation. From the photos, it would appear that the Zen is for businessmen who thought it was a cell phone. The Zen Micro is for stylish women who bought the player to show off that they’re in the 21st century. The Zen Neeon is for women who can’t make up their mind and will buy all 10 colors. The Zen Nano Plus is for the trendy (but not rugged) outdoors types.

With a product lineup like that, it’s no wonder that Creative will need to resort to lawyers to make a buck.

Product Management, Software Business

Ian Landsman has posted a new tour of HelpSpot, and I’ve got to say that Ian has produced a beautiful looking product. I’m going to fire it up to get a first hand look after I move everything over to my new server at GoDaddy, but from the tour this is a great looking product for help desks of all sizes (including 1, as Ian has mentioned).

Technology

Uchi Ogbuji spotted a Slashdot comment by Ian Bicking explaining why TurboGears is not a Rails clone. I don’t think I’ve come out and said it, but what Ian said sums things up quite well. TurboGears is every bit as compelling as Rails, but it’s not a clone. It’s built as an integration of preexisting components, to give you the full front-to-back development stack. We have lots of good tools and ideas to choose from and Rails is not the only idea well from which we’ll be drinking.

Though the parts of TurboGears are not clones of Rails features, David Heinemeier Hansson deserves huge credit for showing what a full stack web application framework could look like at a time when most tools, open source ones in particular, were designed to tackle only one piece of the puzzle.

Python, TurboGears

Ronald Jaramillo has put together CatWalk, a simple model browser for TurboGears. I haven’t tried it yet, but from the screenshots it looks like it could be the start of something quite nice.

Python, TurboGears

A Slashdotting can be very interesting if the thing most slashdotters are looking at is an 80MB file. I have no idea how many people (if any) were served up by the Coral cache. I don’t know how many people downloaded the file via BitTorrent.

I can see a thing or two by looking at my access log, and speaking with Jason I got some other fun facts.

About twice as many people viewed the 20 Minute Wiki screencast today as they did in the previous 23 days. It looks like over 13,500 now. The screencast was using an alarming 80-95Mbps for quite a bit of the day. Apache reports that TextDrive served up more than 1.2 terabytes for me today. Wow! Thanks, TextDrive!

As of now, my server at GoDaddy has served up another 67GB. Bob Ippolito has once again stepped up with capacity, this time on two different servers. Charlie Moad also provided a server. So, the load is now split up among four servers, which helps us all out I think.

As I write this, there are still quite a few people watching, so I’m guessing we’ll tack on another 1,000 views by morning.

Thanks to the folks who kept all of turbogears.org up and available for the whole day!

Python, TurboGears

This is the kind of morning I like to wake up to. A very interesting one. The first thing I noticed was email saying my blog was down. (That’s due to a problem with my MySQL start up script… I guess the server rebooted but MySQL didn’t come back.)

Then I saw the mention that TurboGears has been slashdottted! Thanks to Guillem Cantallops Ramis for getting us on the front page!

And, I was really excited to see a message from Ned Batchelder letting me know that the first TurboGears job posting is now up at Python.org.

Wow. All of this on what is a weekend day for me.

Python, Software Business, TurboGears

Today, we had the first TurboGears Mad Dash. I figured that calling it a “sprint” is probably not accurate, given that most sprints last several times the length of today’s mad dash.

Including myself, we had 10 participants. Given that I had announced my desire to hold a sprint just a few days after TurboGears’ public release, the turnout was very good. Most of the people there were members of michipug (including two that I hadn’t met before), and we also had two hearty travelers from Cleveland.

I thought the day went well! Given that TurboGears is a three-week-old project (built on much older projects, of course), I hadn’t expected that we were going to generate thousands of lines of code (and we didn’t, of course). But, I had expected that people would get a chance to dive in a bit and learn what makes things tick and to collaborate and generate ideas.

On those counts, the day was exactly what I had hoped for. We’ve got some good direction planned for some TurboGears’ 0.9 features, and 9 more people have now had some experience working inside various parts of TurboGears. We also spotted a bug in the forthcoming 0.8 release… it’s always a bonus to be able to stop a bug before the software ships.

A few people expressed interest in doing another sprint, possibly for a longer time than this one (less mad dash-like). The next one will also get moving faster, given that people have more familiarity than they did before. On IRC yesterday, there was talk about a “remote sprint”, which sounds a bit trickier to pull off. We’ll have to see about the remote one, but I’m sure there will be another Ann Arbor in person sprint sometime in the not-too-distant future.

Python, TurboGears