To answer Turgut's points:
>I really mind the dirty talking, 'only 13 sites, and two later on
>declined'
Turgut, this wasn't "dirty talking" but a simple statement of fact. The
first compatibility problem between 1.6 and 1.5o appeared recently, and
someone suggested to solve it by updating the LISTEARN servers. I wanted
to point out that this was not a good solution, because the problem would
disappear from only 11 sites. This was a justification of my decision to
solve the problem by a fix in release 1.6d, which would remove the
compatibility problem regardless of whether or not EARN sites install
fixes from you (such fixes can be sent only to sites which signed the
LISTEARN contract).
>You must have seen the recent changes, like the path-bug and the BSMTP
>work..
To avoid any possible misunderstanding on the part of BITNET readers,
what Turgut calls the 'path-bug' has been fixed (along with another
infinite loop) as part of 1.6d; sites which need the fix can get it from
me (about 5 of them are running it now). I hope to be able to release
1.6d for beta testing at the beginning of next week. I don't know what
Turgut's BSMTP support does as compared to the one introduced in release
1.6a, and, generally speaking, Turgut is much better informed about what
he does than me :-) So, if you have any question about that, please ask
Turgut, not me.
Now I'd like to answer a few of Alain's points:
- (This first point was raised by both Alain and Turgut) I have indeed a
number of beta-test sites, but I lost the "best" ones with the
departure of Jose-Maria (DEARN) and the LISTEARN business (CEARN,
FINHUTC); there are indeed a couple of 1.6 beta-test sites left of
EARN, but low-traffic nodes like ESOC are unfortunately not as useful
as high-traffic ones like UGA or NDSUVM1. Regardless of the
cooperativeness, competence and helpfulness of the people involved in
the beta-test programme, it is always better to test software with
people living in the same time zone as you, and on the same side of a
link where a wrap test for a 1000 records file took over 10h yesterday
(I will not comment here on the reasons why this link is not working
better - refer to the EARNTECH and EARN-NOG archives if interested).
Sometimes I have to wait a full day for a fix to reach the beta-test
site. This is of course better than not having any beta-test site at
all, but this is far from being ideal conditions; if EARN is not able
to organize itself to provide Turgut with beta-test sites, there is a
serious problem.
- Contrary to what Alain seemed to say, I am moving very slowly with new
versions, because I have very little time to work on LISTSERV. In case
you'd doubt it, compare the 'delta' between the shipment of 2 releases
now, and in the early days of 1.5 (up to 1.5n or so). I've been
desperately trying to finish that UDD documentation, but I can only
spend something like 30 minutes a day on it, and I keep being
interrupted, sometimes I have to stop for several days. The result is
that the quality is not very good, I keep having to update it to remove
confusing statements, things are repeated all over the book, and I've
only managed to write about half of it. I should (hopefully) have more
time for LISTSERV within 3 months or so.
- I fully support the idea of introducing formal releases of LISTEARN;
the present situation is quite confusing - some servers have plain
1.5o, some have 1.5o + some fixes from Turgut, some have 1.5o + some
local improvements, some (I suppose?) have Turgut's new version with
the BSMTP stuff, etc.
- Finally, I will not repeat my comments on the NOG's statement that the
directive to sign the LISTEARN contract is purely technical, and not
political. People who are bored enough to want to hear that should
refer to the NOG archives.
Eric
|