| |
In lugnet.robotics, "Brian B. Alano" <alano@kiva.net> writes:
> Since many algorithms apply across languages, what about a single repository
> for all code?
Or perhaps a more general one sitting atop the others (integrationally or as
a launching point). Because even when there's pseudocode for something, the
actual implementation in, say, pbFORTH, may differ fundamentally from the
implementation in, say, NQC.
> (How about lugnet.robotics.rcx.code?)
Do you mean RCX-specific code samples or general LEGO-related robotics code
samples? There are two non-RCX groups (Logo and Handy Board), and possibly
another (Scout) on the way someday, so one notch up under .robotics may be
safer. The higher level area could also be set up as a non-newsgroup
(similar to /cad/datsville/) for collecting information but without the
worry of massive crossposting confusion. I guess there are many options.
> True, it would increase the noise-to-signal ratio for many, but it would
> Keep It Simple.
One nice thing about embeddable pages is that they can be summoned up from
anywhere -- meaning that, for example, a higher-level repository can
aggregate multiple pages from lower-level repositories, making larger, more
encompassing meta-pages.
> Also, I favor follow-ups going directly to the repository news articles,
> for all the benefits seen in the .dat.models Battlestar Galactica example.
> It would also be a good way for people to suggest patches to the originally
> posted code.
I like that approach as well, because you don't have to be using a
newsreader which tracks threads across groups in order to see followups.
The disadvantage is that the group is a little more noisy that way than a
pure announce-style group, but maybe we could add some options to the
news-by-mail gateway and web interface for suppressing followups and/or
highlighting thread roots better.
--Todd
|
|
Message is in Reply To:
17 Messages in This Thread:
  
 
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
