Re: (MS-23) CUSM b9 - Bitch and moan, whine, whine, whine

Richard Cogger (R.Cogger@cornell.edu)
Sat, 10 Sep 1994 08:46:22 -0400


>A few days ago you sent out email regarding the lack of feedback you've
>been getting.
>
>Now, I understand that you have a busy life, what with getting out some of
>the hottest new software on the net, but I have to say that you're not
>using your tight beta-testing community very well, IMHO.

Actually, the beta community is not particularly tight; it's just anyone
who chooses to download and run the beta versions. We do have an alpha
list that is restricted, but nothing has gone to alpha for a couple months
while we struggled with the bugs in 0.70.
>
>> However, these are still *Beta* versions. In particular, we know
>>of one (or more than one) bug that shows up as a crash at startup if the
>>NickNames file has more than 30-some entries.
>
>Since the time that I offered (via private mail) to work with you or your
>developers via mail or over the phone to diagnose the crash I'd been
>seeing, I've sent four revisions of my reflector list to CU-SeeMe-L, which
>you read. It'd have been real nice for you to let us know that you knew of
>this problem. I've sent several messages out saying that I see a crash in
>b7, b8, and would have sent one out about b9. After trimming the Nicknames
>file to 29 names, b9 no longer crashes. Please keep the members of this
>list informed; we're here to help debug your code on a wide variety of
>hardware at no cost to you. Not a bad deal.

Well, Tim heard about this yesterday morning, I think, and I heard about it
about 4PM and announced it at 5:30 or so. How fast do you think we need to
be?
>
>> * Bughalts will now not send you to the dumper (or MaxBug) but will
>>give a dialog with a message and then quit cleanly when you click OK.
>
>Not for me. 070b9 with a Nicknames list > 30 (or whatever my full list is)
>doesn't cleanly exit, it hangs the whole damn machine. I *like* MacsBug -
>it helps me figure out what's going on, tell the developer, cleanly exit,
>and continue doing my work.
>
Well, some bugs lock things up and some trap to MacsBug, if you have it. A
future version will let you edit a resource to switch from the dialog back
to MacsBug. But for many folks who don't run with MacsBug, this change
will let them see the debug message (and report it) and get out cleanly.

>In other news, I requested (via private email) that you or your staff (1)
>arbitrarily decide on a comment character so we can keep contact
>information with the Nicknames, and (2) change the behavior of the Mac
>version so that it not strip from the Nicknames file any lines that aren't
>aliases. I've rec'd no word on either the merit or disposition of these
>requests.
>
A reasonable suggestion; we'll try to get to it.

>There are many people on this list who are using your fine code (thank you)
>for transmitting cultural events around the planet, meeting "face to face",
>and in general attempting to teach a responsible manner of using network
>resources. I believe (and I apologize for speaking for the list
>inhabitants) that the people that I email day in and day out about CUSM are
>ready, willing, and able to be brought closer to the product, even at the
>cost of a few more crashes, hangs, and the like, if we have the support of
>the developers.
>
Well, I'd like to say that we do really appreciate the problem reports and
suggestions we get. They help a lot to refine the software. We do read
them all, at least so far; it may not be possible forever, but we'll do our
best. However, we simply can't respond individually to all of the comments
and reports we get. Many of the messages report problems that have been
already reported or ask questions that have been answered many times
before, even in the Readme files. (This is, of course normal on any list.)
Fortunately, there are some folks like you who patiently answer some of
these queries when they're sent to the list. But we just can't. I do try
to get announcements out about problems that several folks report if it
looks like we won't have a fix soon. Or if we learn of a workaround we get
that out immediately.

>Tomorrow I'll be using both the 68k and ppc versions of b9 to broadcast
>from the Internet exhibit on the floor of the S.F. Exploratorium (from 1300
>to 1700). Bug reports, if any, to follow.
>
All reports appreciated, even if not responded to immediately. To the
extent possible, our response is to try to fix the problems.

Cheers, -Dick