|
|
| Author |
Message |
| 25 new of 123 responses total. |
remmers
|
|
response 25 of 123:
| 
Jun 8 17:59 UTC 2000 |
Re #17: Heh. Turning off the web would be a "the operation was
successful but the patient died" kind of remedy.
(This discussion reminds me that I was responsible for the first
mini-controversy over this issue. A few years ago, before
Backtalk, I wrote a program to convert Picospan items to HTML.
My idea was to use this as a convenient way of archiving old
conferences. When I proposed this in Coop, people objected
because of the search engine indexing issue. At that time I
wasn't well-acquainted with the techniques for preventing it,
and I don't think the objectors were either.)
<end parenthetical aside, back to the subject at hand>
One way to go would be to this on a conference-by-conference
basis, with the decision left to the conference participants,
and in no case index older material retroactively.
|
pfv
|
|
response 26 of 123:
|
Jun 8 18:11 UTC 2000 |
Actually, I would think that:
1) nothing before acceptance was indexable;
2) Items should be flagged as indexable and ANNOUNCED at reply;
3) The current "Ok to enter this response?" garbage would go
another step - "You know this item is web-indexable?".
Personally, I see absolutely zero advantage to grex OR mnut ever
being as stupid as to allow their confs/items/resp being
indexable. HOWEVER - I can see a local "index" that does a
"find.." on a grand-scale. particularly from the webside, since
the concept of a "hyperlink" in picospan is impossible.
I can even see such an invaluable capability being "member-only"
in hopes of gaining even further fundage..
|
scg
|
|
response 27 of 123:
|
Jun 8 18:12 UTC 2000 |
My guess is that such a change would have a pretty serious chilling effect
on the conversations here. I'd certainly think very carefully before posting
anything here that was likely to become the first thing anybody would find
out about me, doing a search of the entire web.
Now that I think about it, though, letting the web search engines loose on
the conferences would have an effect far beyond making the conferences easily
searchable. If somebody wanted to find actual information about anybody who
posted a lot of stuff on Grex, or anybody whose name was used as a pseudo by
somebody who posted lots of stuff on Grex, the volume of the Grex conferences
would be massive enough that finding the real information would be pretty much
impossible.
|
pfv
|
|
response 28 of 123:
|
Jun 8 18:30 UTC 2000 |
The latter does absolutely zilch to reassure me.
|
gull
|
|
response 29 of 123:
|
Jun 9 19:54 UTC 2000 |
I don't like the idea, simply because I've seen what a deep search by a web
spider does to a web site. Even indexing of a relatively simple site will
almost totally shut that site down for a while, if it's on a 56K modem. I
think the affect of a web-bot indexing thousands of responses would be
devestating to Grex's bandwidth.
|
mdw
|
|
response 30 of 123:
|
Jun 10 00:17 UTC 2000 |
The smarter indexing engines pace themselves so they don't chew up all
the bandwidth to a site.
|
janc
|
|
response 31 of 123:
|
Jun 10 04:13 UTC 2000 |
(After all, if you are indexing the whole web, you can just as well work
on 2000 different sites at once, so each one is being hit by only one in
2000 of the queries you make, so no one site is being hammered hard.
It's not hard to do, costs you little, and avoids getting you hated by
every site on the net. I'm just assuming that reputable search engines
would do something like this.)
|
swa
|
|
response 32 of 123:
|
Jun 12 03:13 UTC 2000 |
Cmcgee writes in #6, " In fact, part of my willingness to post on Grex has been
premised on the idea
that no one could find my thoughts unless they found Grex first and also were
interested enough to join a conference."
Yes. Me too. I do recognize that Grex is already, as lots of people have been
pointing out recently, a "public forum." But it is also a community, and I
think that making it *more* public in this manner would be a big mistake. So
much of what is posted here is not just factual information but personal
reflections, life stories, emotions... I think that making it easier to find
all this via a search would do far more harm than good.
|
aruba
|
|
response 33 of 123:
|
Jun 12 03:29 UTC 2000 |
I agree that I'd prefer people not find responses out of context. Whole
items I maybe could be talked into, but I dunno.
|
jep
|
|
response 34 of 123:
|
Jun 12 17:10 UTC 2000 |
I'd love to see a Grex-specific search engine, so that I could search
for information I've seen people enter before on Grex.
|
cmcgee
|
|
response 35 of 123:
|
Jun 12 20:42 UTC 2000 |
Now THAT would be a neat trick. Saves having to estimate when I might have
seen it and searching every conference along with selected agoras.
|
remmers
|
|
response 36 of 123:
|
Jun 12 22:01 UTC 2000 |
We could run the 'glimpse' search engine here, if we have the disk
space. (Glimpse builds rather large index files.)
|
janc
|
|
response 37 of 123:
|
Jun 13 02:52 UTC 2000 |
(Backtalk and Picospan both have search functions that work within a
single conference - neither one is blazingly fast since they don't use
indexes.)
|
jep
|
|
response 38 of 123:
|
Jun 13 15:20 UTC 2000 |
Is disk space really that much of an obstacle? Why?
|
goose
|
|
response 39 of 123:
|
Jun 13 18:36 UTC 2000 |
glimpse searches would be great here.
|
hhsrat
|
|
response 40 of 123:
|
Jun 14 02:22 UTC 2000 |
Re #38, 39 - I (my own opinion, not reflective of Grex Policy, or the
opinions of anyone who knows what they're talking about) would like to
be able to search, using indexed searchers or whatever, and think that
this could probably be done without a need to add additional disk.
However, I'm sure that pfv would be willing to post a long list of
eggdroppings that he sees as an obstacle to disk space. (this is not an
invitation)
|
janc
|
|
response 41 of 123:
|
Jun 14 04:42 UTC 2000 |
Backtalk doesn't do indexed searches because they are harder to program and
I'm not highly motivated. Disk space could certainly be found for indexes
if we were highly motivated.
|
mdw
|
|
response 42 of 123:
|
Jun 14 06:52 UTC 2000 |
PicoSpan didn't try to index anything because (1) in 1983 disk space was
much more expensive, (2) it's surprisingly hard (and somewhat slower) to
properly update a dynamically allocated data with variable sized text
and keep information in several different places in sync, even in the
face of problems like an untimely process death, system crash, or
hardware (ie, disk) mis-behavior.
The other problem with indexed text is, um, how *do* you index english
text so that it's possible to get quick results, do various sorts of
wildcard searches, and not consume gobs of disk or CPU? I've been kind
of curious just how the web search engines tackle these problems, but
haven't seen any good description.
|
jp2
|
|
response 43 of 123:
|
Jun 14 11:46 UTC 2000 |
This response has been erased.
|
remmers
|
|
response 44 of 123:
|
Jun 14 16:42 UTC 2000 |
Or try the book _Managing Gigabytes_, by Witten, Moffat, and Bell
(Morgan Kaufmann 1999).
|
johnnie
|
|
response 45 of 123:
|
Jun 16 01:36 UTC 2000 |
So what are we imagining people would be searching for to run across
Grex items? Names? General topics? I just did a search for " 'John
Hofmann' and Grex " and came across
http://www.grex.org/grexdoc/archives/minutes/1993-11-22, which was news
to me 'cuz I didn't know I'd ever attended a Grex Board meeting. Hmm.
I also remember doing a search a couple of years ago on my name only and
pulling up an agora response of mine where I called my then-boss a
doofus or somesuch (that's when I stopped using my name as a psuedo).
|
remmers
|
|
response 46 of 123:
|
Jun 16 13:10 UTC 2000 |
Hm, it seems I was secretary then and took the 11-22-93 Board minutes.
Well, looking up the minutes, it appears that someone claiming to be
John Hofmann and have login 'freedog' was among the attendees...
As for what people will want to look up, I'd guess topics would be
a major one. For example, I might be in the market for a new car
and remember that there was a discussion on the relative merits
of various local dealers a while back. Can't remember if it was
in the Agora or Cars or Consumers conference, or just when the
discussion took place. Being able to search across all of Grex's
conferences on key words and phrases like "cars", "dealerships",
and "Ann Arbor" would help me track it down.
|
scg
|
|
response 47 of 123:
|
Jun 16 17:59 UTC 2000 |
I don't have any objection to being able to search within Grex. That strikes
me as very different than searching the web and finding a Grex discussion.
|
remmers
|
|
response 48 of 123:
|
Jun 16 20:21 UTC 2000 |
It is quite different, and would serve different purposes.
If people have qualms about across-the-board access to Grex
discussions by web search engines, then I think it would be
reasonable to make it an option on a per-conference basis, with
the decision left up to the conference participants.
A big plus for making some Grex discussion web-searchable is that it
could make Grex more visible on the web and draw in more participants.
This might help counter the "same people always saying the same stuff"
syndrome, and could be quite healthy for Grex.
|
johnnie
|
|
response 49 of 123:
|
Jun 18 02:49 UTC 2000 |
re 46: Yep, "freedog"--'twas I.
|