comparc@twwells.uucp (comp.archives) (11/21/88)
I have been receiving most non-database comp.archives messages as
other than messages to comp-archives. So, I've been posting most
messages in this more-or-less digest form.
I could make it official and post everything other than the database
messages in digests. If I did so, I'd have to have someone tell me
what the rules are for making a digest. (Should I follow the form
used in comp.sys.sun?) Does anyone have any feelings about this, pro
or con? Any information?
---
Here are a couple of changes to the database format:
1) The definition of a site is somewhat vague. What I am going to
do is to consider one set of archives under the control of a
single administrator as an archive site. This means that the
site entry won't have different sets of data for archives
located at the same site. This also means that the archive
name will be somewhat less related to the address of the
archive.
2) The access method and access tag of the CO fields have been
swapped. The access method now comes first.
---
prapc2!paulr (Paul Raulerson) sends:
: I opened up one of my machines over here for public access, mainly because
: there is a lot of intrest in UNIX over here from folks that have zero access
: to a UNIX machine. It has taken off like a banshee, with between 4 and 15
: people per *day* asking for access to this machine. (*sigh*)
:
: Obviously, with only one dial-in line to this machine, I cannpt support
: that kind of user load, but since this support is coming straight out
: of my pocket, my better half (and I agree with her) says we cannot afford
: to pay for more phone lines or more telecom equipment or just about any-thing.
:
: I have been considering charging a monthly fee for access to private lines,
: and leaving the one public line open. In your opinion, would that
: consistute any breach of usenet ethics? In particular, I would like
: to charge enough ( say $3 to $5 per month per person) to allow me to
: put some archival storage on line. I don't think I want to do this as a
: business, as it would almost certainly then preclude most of the
: kids around here. (Some of them are pretty darn brilliant!)
Well, remember that Usenet is very much a voluntary and libertarian
(dis)organization: "commercial" is what someone calls commercial.
So, while there are some things that no one is going to argue about,
there are some grey areas. One of them is the kind of site you are
describing. Some people are going to look at it and say: "He's
charging money! Commercial!" Others are going to say "But he's not
making a profit, so he's not commercial." However, I imagine that the
bulk of Usenetters will greet this with a great big ywan, so you
should be safe.
Anyway, I think what you are doing is a Good Thing, and certainly
non-commercial, so you won't hear any complaints from me.
---
datapg.mn.org!sewilco (Scot E Wilcoxon) sends:
: I'd prefer if comp.archives only posted archive info. I'm trying to
: archive the group (for later building an index, when we have the tools),
: and am already having trouble with a lot of text messages in the group.
To save yourself headaches and disk space, only save those entries
whose subjects begin with 'DB:'. Those are the ones that contain the
information you want.
---
june.cs.washington.edu!pardo (David Keppel) sends:
: Lately you have been posting articles verbatim and then
: following up with your own comments in a separate message.
: This requires including parts of the original text. In the
: standard spirit of a moderated newsgroup, how about just
: inserting your editorial comments in brackets?
Ok. I had thought that was only done in digested newsgroups, so I had
been avoiding putting editorial comments in the messages themselves.
---
jetson.UPMA.MD.US!john (John Owens) sends:
: > comparc:comp.archives::comparcupd /usr/home/comparc/db
:
: Actually, news 2.11 checks both the newsgroups and the distribution
: against the second field of the sys file line, so it needs to be
:
: comparc:comp.archives,world::comparcupd /usr/home/comparc/db
:
: or even ",world,na,usa" if you imagine restricting archive postings
: geographically.
Thanks. That saved me the hassle of figuring this out for myself. As
for restricting the postings geographically, I can't imagine any
reason for doing so.
: About your From: line problems - save your .newsrc and try vnews, just
: as a test. If your From: lines have domains and your 'r'eply command
: uses the domain address, substituted into your "internet" line, then
: it's an "rn" problem. Since I know nothing about rn, I can't help you
: there, but if it doesn't show the domains in vnews, I can try to
: figure it out.
I get to be red in the face for this one. The problem has nothing to
do with news. You see, when you post, the message gets MAILED to me.
And, of course, my mailer is going to put the appropriate paths in
the mail header. What this means is that I get to figure out how to
turn a mail header into the desired From: line. Does anyone know
where I might find some hackable code to do this?
---
uokmax!kwthomas sends:
: The site that I'm logged into is both an internet and a uucp site. When
: I sent the letter out, I used the "user@site.uucp" notation. The mailer
: took care of the rest. The internet machine name is "uokmax.ecn.uoknor.edu"
: while the uucp name and reasonable address is:
:
: att!occrsh!uokmax!kwthomas
Well, I happened to use the uucp address, but on second thought I
should have used the Internet format. From what I've seen, it is
easier for a uucp site to get to the Internet than it is for an
Internet site to get to uucp. Not only that, but there is something
of a push to get away from uucp style paths and into domain names.
What really threw me was the From: line:
: From: proxftl!gould!sun!Central!uokmax!kwthomas%uokmax
From what I understand, this is wrong. You should probably check with
your sysadmin to see if this is what is supposed to go out. I would
have expected something more like:
proxftl!gould!sun!Central!kwthomas%uokmax.ecn.uoknor.edu
or
proxftl!gould!sun!Central!uokmax.ecn.uoknor.edu!kwthomas
: The "path" at the top of the news article looks correct.
The Path: line is generated by the news software as it passes through
each machine; I have nothing to do with it, other than to see that it
gets started right.
: As to the "message id" line, I don't know where that one
: came from.
That comes from your system. Many mailers generate a unique message id
for each message that you send. This makes keeping track of things
easier all 'round. I preserve those whenever I get them.
---
buff@pravda.gatech.edu (Richard Billington) sends:
: I'd like to suggest that a collection of FTP-able
: files be made available that look like this:
:
: database-core.ddmonyy
: database-update.ddmonyy
: .
: .
: .
:
: Where the database-core is always the most current
: version of the database, and the update files are
: for those who've been maintaining the database all
: along by getting the monthly update files and adding them
: in.
Anyone running an ftp archive can do this if they want to; however, it
is not sufficient. Ftp is not the whole world; there are plenty of
sites that can't do ftp. Like mine, for instance.
: Your news postings would then only have to take up the
: summaries of what had been posted. This seems like it
: would cut down on news traffic (always desireable) and
: make it easy for people to (a) start using the database
: whenever they wanted and (b) make it easy to catch up,
: because I should the speed of ftp over getting it off of
: rn is much greater (c) make it easier to follow what
: additions were made because there'd be an easy to follow
: summary on news.
I could post summaries instead of the whole thing, assuming that
there are archive sites set up to contain the databases. However,
there is not, yet, sufficient volume to go this route. And, of
course, no such archives exist. When and if, I'll look into it.
: If some people need to have rn as the distribution media
: for the database, then how about a seperate news group
: that is just database updates - moderated to insure that's
: all that get's put in there.
Actually, the newsgroup proposal I preferred was to have both a
moderated newsgroup and an unmoderated newsgroup, the former devoted
to archive postings and the latter for discussions relating to
archives. That was not the idea that got voted in, so we have one
newsgroup.
---
pcrat!rick sends:
: In article <8811120526.AA01398@uunet.UU.NET> you write:
: > I had to make a few changes to your entry.
: >
: > All addresses should be user@site and were made so.
:
: Thanks for most of the edits. However, these @ addresses are
: not only impossible to read, but they don't work for most people
: not on the internet.
It very much depends on which system you are using. Our Sun (SunOS
4.0) understands them. So does any machine running smail (which,
unless someone trashed the copyright notices on my copy, is public
domain).
: In fact, I've *never* gotten an @ address
: to work. A bang path always works, I just mail to uunet!site!user.
: The least common denominator is bang paths, so why not keep them?
As it happens, I also always use bang paths, having gotten used to
them when I didn't know about the domain name syntax. However, I
intend to switch over, just as soon as I get my maps set up.
Uucp style addresses are necessary when the network you are using
requires you to tell it the route for delivering your message.
However, such networks are a real pain, requiring every user to
become something of a network guru. The direction things are moving
towards is to have you specify where you want the message to go and to
let the software figure out how to get it there. The domain names are
a part of that move and so I'll stick with them.
BTW, did you know that your message had two copies of your signature?
---
microsoft!w-colinp sends:
: A couple of comments:
:
: I don't understand the time of day format. Could you explain it?
: My idea of a nice format is:
:
: - No time specified means "any time". This is pretty common, and
: it doesn't seem worth wasting space storing this.
That's the way it works.
: - Days of the week are specified MTWRFAU, with thuRsday, sAturday, and sUnday
: hacked up. Also add H for holidays.
I'm using two letter abbreviations. These are better known. As for
holidays, what qualifies as a holiday isn't well defined, so I'm not
going to try to deal with it.
: - Times of day may wrap past midnight. E.g. I can say MTWRF1900-0600
: to say that weekday nights is okay. Be careful not to wrap past your
: start time!
This is allowed.
: - Question: should multiple weekday specifications be colon-separated? The
: format doesn't need it, but it may be nice.
They are to be separated by semicolons.
: Also, some description of the form of the file might be nice.
: E.g. ASCII text, C source, WEB source, Mac binary, C source patch,
: Bibtex, etc. This is sort of covered by your tools section, where
: you can list uudecode, uncompress, patch, arc, zoo, cc, or whatever.
: No tools, of course, means ASCII text (e.g. purity test), but how do you
: say "needs ANSI C compiler"?
That is covered elsewhere. There are actually three databases, though
we are concentrating on just one of them right now. One of them (the
info database) describes what an archived item is, and can include
information on the compiler, etc. needed.
: One more thing: a standard entry for mail servers (e.g. netlib) would
: be nice. I recall you mentioning that you haven't had much experience with
: these, but they're pretty easy to use - just send mail with subject
: line "help" to netlib@research.uucp or netlib@anl-ncs.arpa.
: This will explain how the system works. You can get the netlib software
: from these sites, too. Might be useful in your own setup.
The problem with mail servers is not how easy they are to use, it is
that I am not sure how to standardize their entries.
---
gnu@toad.com (John Gilmore) sends:
: I thought the purpose of comp.archives was to post listings for source
: archives.
Well, the purpose of comp.archives is to let people know how to get
to archives, source or not.
: And I thought it was moderated so we wouldn't get trash
: like the above. [a random looking-for message. tww]
It is moderated so that the archive listings, etc. can be
distinguished from the conversations, and to keep the conversations
related to archives. I have a good reason for permitting those types
of entries: I want people who know where these things are archived to
let us know. Perhaps, once the listings are more complete, I will
dispense with those messages. In the mean time, they serve a purpose.
: So far all I see is discussions about how great the
: newsgroup is, and cryptic entries with no explanation of what they
: mean.
Well, the explanations are available. See messages <160@twwells.uucp>
and <161@twwells.uucp>. Also, you might have noticed that I am
posting a slightly less cryptic interpretation of the data. It's not
perfect, but I'm working on it.
: If people send you random requests, redirect them to comp.sources.wanted,
: which is where the one above should have gone.
:
: Discussions of source related stuff should be in comp.sources.d.
Like I said, I'm only taking the "where is an archive that has..."
messages. Anything else gets trashed.
: PS: "twwells" is not exactly the world's best known uucp site. Can
: you put a few known paths to it in a signature in each message you post?
: Not everyone uses "routing" mailers, and the maps are always broken anyway.
I put a signature in every message I originate; it includes a path
from uunet. Are you suggesting that I should put a signature in every
message, even those that someone else originates? Also, as soon as
Telebit gets me my modem, I'll be directly connected to uunet.
---
datapg.MN.ORG!sewilco (Scot E Wilcoxon) sends:
: I suggest you keep the newsgroup alive by posting the entire DB
: regularly. When the software has been distributed and is working,
: you can then cut the entire DB back to posting chunks of it
: to keep everyone up to date.
This sounds like a good idea. So, for now, until the databases get
too large, I will repost them with the other messages I regularly
send.
: You'll need to repost everything
: regularly for newcomers, just as comp.mail.maps works.
Actually, I was hoping to be able to post instructions so that
newcomers can get the current set of databases from some archive site.
This would make it unnecessary for me to repost the entire database.
We'll see.
---
Bill
{uunet|novavax}!proxftl!twwells!bill
send comp.archives postings to twwells!comp-archives
send comp.archives related mail to twwells!comp-archives-request