If you’re trying to build a Python package (in my case, the ReportLab toolkit), and your build fails with a string of error messages starting with Python.h: No such file or directory, the problem is that the package includes some amount of code which is written in C. The build is trying to compile that code, and the C compiler is looking for the Python C headers, and for most Linux users (I ran in to this on an Ubuntu system) the C headers aren’t part of the core Python package. You need the python-dev package. Try this:

sudo apt-get install python-dev

Then try your build again; I bet it will work.

We aren’t lawyers–that’s Scott’s area–so I won’t attempt to give a detailed explanation of the application’s appeal to corporate legal types who might see non-disclosure agreements crossing their desks in droves. We were interested in the technical challenge: could we build a web application that could be trained to “read” a Word document and recognize about 28 common clause types? Could we then alter the document according to a flexible set of rules, and return the altered document to the user?

We think we’ve done that, with Scott’s guidance, and late last week Ken Adams of AdamsDrafting.com had some pretty good things to say about the system as well. If reading and vetting NDAs is a part of your responsibilities, take a look at BaselineNDA; you might find it useful, and that’s the highest praise you can give us. (If NDAs aren’t part of your responsibilities, take a look at the demo; that might explain why we’re so pleased with the thing.)

If you make a spelling goof–or, indeed, nearly any other goof–in a message you’ve sent in Skype’s IM window, you can fix it using Perl’s substitution syntax. In other words, if I was to type

I sent it to you're work email

and commit that, I could then type

s/you're/your/

(which means, in Perl, “substitute the second term wherever you find the first”) it would correct my earlier gaffe, and change the timestamp to indicate that I’d edited the message.

This only appears to work in the most-recent message sent, of course.

Here it is in action:

Technorati Tags: perl, regular expressions, skype

If it wasn’t already obvious that these companies were bottom-feeders, their data entry skills would tip us off. Lately there has been a wave of junk mail based on the erroneous idea that our company is “Common Medical, Inc.”

One hopes this won’t progress to the point where we’re getting pharmaceutical samples in the mail.

But it’s harder than we realized. Things will get easier, I suspect, when we have an actual site we can share with you; right now, my gut feeling is that talking about the site is promoting buzz about “vaporware.” When it’s real, we can talk about it, you can see what we’re talking about, and this weblog can be a conversation between you and us about how we can improve it.

Even the development process is hard to talk about. As you may have figured out, we’re working with Ruby on Rails, which has dramatically improved the time it takes to get us from ideas to HTML in the browser; both Noah and I have extensive experience in LAMP, and working in this framework is positively eye-opening. But we’re new at it, no question. We discover new things every day, every hour, and we spend a certain amount of time going back and re-doing things we’ve already done because we’ve discovered a way to do them better.

Discussing the details of every new discovery, things everyone else who’s been using Rails for months longer than we have already knows, doesn’t seem like a way to engender confidence in our users and potential investors. It’s scary stuff.

Instead, I find myself writing 300-word micro-essays as comments on tickets in our internal Trac system. (Trac, by the way, I am in love with. Have I mentioned that?)