f/f george wrote:

On Sun, 25 Jul 2004 17:22:37 +1200, "~misfit~"
<misfit61nz@yahoo-mung.co.nz> wrote:

f/f george wrote:

On Sun, 25 Jul 2004 02:20:21 +0100, Barry Walsh
<bwalsh@SPAMMERSAREVERMINindigo.ie> wrote:

In the case of Predictor they recently forced everyone to upgrade
to
3.20 to continue receiving work as it fixed a bug that was screwing
up a number of their workunits. The Predictor website also had the
3.20 client a couple of days before Berkeley themselves listed it.
Inserting a version notification system might be a good idea. It
seems to check the version anyway, but presenting that info to the
end-user would help. So far the only way is to manually check the
BOINC website or wait until your client is too obsolete and an
upgrade is forced.

PERSONALLY I think that EVERY upgrade should be mandatory! BUT I DO
realize that not everyone has time to do all of them such as those
that have many machines like in a business environment.
What kind of problems can supporting several different versions of a
piece of software wreak on programmers? I would bet alot! Keeping
old code just so some people won't have to upgrade when the newer
version of the software fixes or provides better support for some
function.

Why can't it autoupdate itself? I just did another install of 3.20
on a machine and the first file it downloaded was an .exe of 1.5MB.
Then it managed to grab *one* WU. That's all it's got in 24 hours.

It COULD, but the people that run Boinc in a business setting have
complained that they want to check it out first, just to make sure it
doesn't mess up their business.

Yeah, I guess I understand that.
--
~misfit~