January Weiner wrote:
> Why didn't you put your package [EMBOSS] into the freshmeat
> database?
I've noticed that very few of the chemistry and bioinformatics
software projects are listed in that database.
26 hits for "molecular"
2 for "bioinformatics"
8 for "protein"
1 for "nucleotide"
about 5 for "DNA" (lots of false positives)
given the duplicates, this means about 35 programs for
this fields in that database. I can name at least
that many more programs in common use.
I suspect there are several reasons for this:
1. People learn about projects from reading papers, from
friends/colleagues, from attending conferences, from
newsgroups and from web searches.
2. Because few bio/chem projects are on freshmeat, few
people look to it as a resource and instead use 1.
3. Similarly, most developers use methods from 1. to
learn about projects so don't think to go to freshmeat
as an announce site.
4. I personally used to read freshmeat almost every day to
see what was new. Now I almost never use it because
there are too many projects listed, so finding something
relevant is difficult unless you already know the name.
(Eg, search for "RNA" and there are more than 400
results, like "jou-RNA-l"). And if you know the name
its easy to find the project.
5. Because I don't use freshmeat, I don't think of it, so
I make my annoucements on the relevant newsgroups and
mailing lists (bionet.software, CCL, biopython, etc.)
6. Quite a few of the packages available are mostly packaged
forms of some internal work, perhaps to go along with
a paper publication. Making a program which is portable
and usable by others is much harder work, and doing
the work to announce a project more widely (like with
freshmeat) requires effort both to make the annoucements
and in support. Many people likely don't want to go
through that hassle, especially if they aren't using
the code any more in-house.
Thoughts-on-a-sunny-but-cold-winter-day,
Andrew
dalke at acm.org