Prev: regression through a fixed point
Next: Build EX Problem
From: Walter Roberson on 28 May 2010 22:12
sscnekro wrote:

> 2. Separating Matlab news from Simulink news.

When the newsgroup really gets going around October and March, it can
get over 500 postings per day. That volume is a problem to keep up with,
and at times like that it would be useful to have finer-grained
categories. But at this time of year, there are less than 150 postings
per day, and there can be more than an hour go by without any postings
at all, not nearly as hard to keep up with.

It is plausible that some people are daunted by the volume and would
feel more comfortable replying in smaller venues -- but that's speculation.

I tend to learn a fair bit here "accidentally", reading postings that
don't have anything to do with my work or experience.
From: sscnekro on 28 May 2010 22:24
Hey guys,
it's perfect if you comment on feasibility & technical limitations to make any of such changes to the newsgroup and even predict people's behavior. The question was a bit easier - just basically to find out whether you'd appreciate should some of those things work. As for me, as long as I get hlp with ML problems, no matter how much messy the site. On the other hand, I doubt it'd be that difficult to make the site more structured and attractive even though keeping the character of posts as plain as possible.
From: sscnekro on 28 May 2010 22:39
> Re: Walter Robertson [I tend to learn a fair bit here "accidentally", reading postings that don't have anything to do with my work or experience.]

Mhm, I'm trying to read most of the new posts too, but as a basic ML user, find it often difficult to relate the posts to some more general topic. I think this is a perfect environment to learn from others .. I could imagine e.g. there'd be list of about 10 themes when first posting (more than one could be ticked [x]) that would then appear as the first lines of the field. I think the concern of some that a bit of more topic fine-tuning 'd reduce the informativeness and variety of what you follow, is ilusionary. In fact, it could be the opposite.
From: dpb on 29 May 2010 00:04
sscnekro wrote:
>> Re: Walter Robertson [I tend to learn a fair bit here "accidentally",
>> reading postings that don't have anything to do with my work or
>> experience.]
>
> Mhm, I'm trying to read most of the new posts too, but as a basic ML
> user, find it often difficult to relate the posts to some more general
> topic. I think this is a perfect environment to learn from others .. I
> could imagine e.g. there'd be list of about 10 themes when first posting
> (more than one could be ticked [x]) that would then appear as the first
> lines of the field. I think the concern of some that a bit of more topic
> fine-tuning 'd reduce the informativeness and variety of what you
> follow, is ilusionary. In fact, it could be the opposite.

I disagree (once I finally figured out what you're trying to say --
what's up w/ the "'d" thingie??? that's not a valid contraction as I
noted earlier; it just confuses/or at least muddles the query.

Anyway, I basically read 'em all as they come as Walter notes; if I
don't care and don't know anything useful (or don't have time to spend
at the instant even if do) I go on; occasionally there is some tidbit
that is intriguing even if not directly in my line so I read on.

OTOH, the SIMULINK queries rarely are of interest for topic and unless
they're generic Matlab I could safely just filter them and not miss them
I think.

I still think other than for product-based at most there really is no
sense in trying to make more than there is.

As for the interface, that's immaterial to me; as noted earlier, it took
over a minute to load a single thread from the TMW web interface whereas
I can download multiple groups headers in that time. That's a killer to
even think of adding more.

--
From: Walter Roberson on 29 May 2010 22:26
sscnekro wrote:
>> Re: Walter Robertson [I tend to learn a fair bit here "accidentally",
>> reading postings that don't have anything to do with my work or
>> experience.]

Do you have an email address for Walter Robertson? He's taking credit
for things I wrote, and that must stop.
First  |  Prev  |  Next  |  Last
Pages: 1 2 3 4
Prev: regression through a fixed point
Next: Build EX Problem