Lambda the Ultimate

inactiveTopic Sociology of Language Development (Python)
started 2/9/2001; 6:29:48 AM - last post 2/13/2001; 1:25:55 PM
andrew cooke - Sociology of Language Development (Python)  blueArrow
2/9/2001; 6:29:48 AM (reads: 2761, responses: 5)
Sociology of Language Development (Python)
I think this thread shows how tricky it can be to keep the balance between developers and users of a language. IMHO Python has got it just about right, with PEPs describing new features and (until now) regular updates on the development process posted to the c.l.python.

Even so, there's clearly some friction. If anyone can compare this with Perl development, I'd like to know if they do any better.
Posted to Python by andrew cooke on 2/9/01; 6:30:26 AM

andrew cooke - Re: Sociology of Language Development (Python)  blueArrow
2/9/2001; 11:16:34 AM (reads: 1284, responses: 0)
The story continues.

Ehud Lamm - Re: Sociology of Language Development (Python)  blueArrow
2/9/2001; 1:54:28 PM (reads: 1273, responses: 0)
How about other language revision processes? I pointed to some Ada documents from the design documents page, and many other documents are available. The procedures for producing 'change requests' are here, but the top page is a better starting place. Notice, that though the page says the ARG deals mainly with errors in the standard, there are AI that deal with major language enacncments (like adding multiple interface inheritance).

Does any one know of (academic) work on various language evolution processes?

Chris Rathman - Re: Sociology of Language Development (Python)  blueArrow
2/9/2001; 4:31:16 PM (reads: 1253, responses: 0)
Don't know how things are now, but when Larry Wall announced an anything goes policy for the next version of Perl 6.x, there was a tidal wave of activity on the mailing lists. A gazillion suggestions, from the minor facelifts to what would be a complete new language.

The dialog was not too productive because there was no way to filter out the good and bad, as well as no way to pursue lengthy discussion since no one knew what ultimately be adopted. IIRC, there was some brouha when a few of the main developers decided to take the discussion to a quieter (but restricted) mailing list, as some felt left out of the decision making process.

My feeling is that no matter what language you're dealing with, there has to be someone (or some committee) in charge to make the decisions. Anarchy is the result if you don't have that filter in place as everyone has things they want to change - but it's rare that the changes actually overlap from developer to developer. Any language revision process needs to weigh the advantages and disadvantages of change.

Ehud Lamm - Re: Sociology of Language Development (Python)  blueArrow
2/13/2001; 11:05:30 AM (reads: 1227, responses: 1)
Of course this is important. I just read a paper on the history of Algol-68, and this is made very explicit there. I pointed to the Ada process, since the documents are available online, and since I think the process was very well organised (see the mapping documents, lanugage study notes, rationales etc.).

Related to this, see the book The Evolution of C++, Language Design in the Marketplace of Ideas edited by Jim Waldo. This book deals with specific technical issues (e.g, multiple inheritance), and is collection of papers outlining suggestions and critiques. It shows one phase of the process, the fermentation.

Aside from papers about specific language histories, I was unable to find a general work about the sociology of language development and use. I am sure they are out there.

Ehud Lamm - Re: Sociology of Language Development (Python)  blueArrow
2/13/2001; 1:25:55 PM (reads: 1295, responses: 0)
I am sure they are out there.

And if not, someone should probably go ahead and write on this topic.