how badly broken is google's usenet archive?
James Dolan
google's usenet archive?
i've often heard people complain about it, but i've usually found it
to be not as badly broken as the worst complaints seemed to suggest,
until i tried using it this week and found it to be just about
completely dead. is it true that it's just about completely unusable
now? about how long has it been that way?
i apologize if i'm repeating questions that people are already sick of
answering, but because of the archive problems i haven't even managed
to find out to what extent the problems have already been discussed.
--
Mensanator
> my basic question is in the subject header: how badly broken is
> google's usenet archive?
Not bad.
>
> i've often heard people complain about it, but i've usually found it
> to be not as badly broken as the worst complaints seemed to suggest,
There will always be whiners who blame the system
for their own shortcomings.
> until i tried using it this week and found it to be just about
> completely dead. �is it true that it's just about completely unusable
> now? �
No. I just did a Google search and found the very
first message I ever posted from Jan 6, 2001. Couldn't
do that if it was broken.
> about how long has it been that way?
You're asking for a conclusion based on a false premise.
Bill Dubuque
It appears to be badly broken. Many of my posts have links to
prior posts (using the message-id) and many of these links no
longer work (though they all used to). Many keyword searches
no longer find posts they used to. Attempting to browse the
prior posts of someone via "View Profile" often yields links
that give errors or far fewer posts than the displayed count.
Etc, etc. These problems appear to have been introduced in the
last half year or so. Coincidentally Google Groups acquired
a new manager in that time frame (according to a post in the
Google Group support forum). It's a crying shame that we no
longer have a reliable archive. So much for Google's motto
"do no evil".
Chip Eastham
I noticed that some of the summary stuff,
indexing of recent posts by author, was
broken over the last couple of months,
but I think this got caught up in the
last week or so. Google Groups does
have folks who respond to reports of
broken functionality, though I imagine
not everyone hears back about specific
fixes.
regards, chip
William Elliot
> my basic question is in the subject header: how badly broken is
> google's usenet archive?
>
As badly broken as your punctuation.
Phil Carmody
> my basic question is in the subject header: how badly broken is
> google's usenet archive?
Google groups sucks.
> i've often heard people complain about it, but i've usually found it
> to be not as badly broken as the worst complaints seemed to suggest,
> until i tried using it this week and found it to be just about
> completely dead. is it true that it's just about completely unusable
> now? about how long has it been that way?
I think while they were just replicating DejaNews, which was great
before they sold out and became a 'portal', things were pretty OK.
I suspect they changed primary key at some point, and they bollocksed
up their joins so searches simply stopped working.
My favourite demonstration that google groups sucks is to simply
search for "Google groups sucks" (in quotes). If you add my surname
to the search, you'll find an explanation why as the first hit, but
that explanation won't be found if you do not include my surname.
So a subset of the results contains something which wasn't in the
original results. That's simply broken. Google groups sucks.
> i apologize if i'm repeating questions that people are already sick of
> answering, but because of the archive problems i haven't even managed
> to find out to what extent the problems have already been discussed.
It's not really a sci.math question - try alt.google-sucks perhaps?
Phil
Crosspost & Followup-To: alt.google-sucks
--
Marijuana is indeed a dangerous drug.
It causes governments to wage war against their own people.
-- Dave Seaman (sci.math, 19 Mar 2009)
James Dolan
<w...@nestle.csail.mit.edu> wrote:
|jdo...@math.UUCP (James Dolan) wrote:
|>
|> my basic question is in the subject header: how badly broken is
|> google's usenet archive?
|>
|> i've often heard people complain about it, but i've usually found
|> it to be not as badly broken as the worst complaints seemed to
|> suggest, until i tried using it this week and found it to be just
|> about completely dead. is it true that it's just about completely
|> unusable now? about how long has it been that way?
|>
|> i apologize if i'm repeating questions that people are already sick
|> of answering, but because of the archive problems i haven't even
|> managed to find out to what extent the problems have already been
|> discussed.
|
|It appears to be badly broken. Many of my posts have links to prior
|posts (using the message-id) and many of these links no longer work
|(though they all used to). Many keyword searches no longer find
|posts they used to. Attempting to browse the prior posts of someone
|via "View Profile" often yields links that give errors or far fewer
|posts than the displayed count. Etc, etc.
hmm. what you're describing doesn't sound anywhere near as bad as the
way that google groups searching has been working for me in the past
week. i pretty much can't get _any_ usenet searches to work at all
via google groups.
(i tried phil carmody's suggestion to search on "google groups sucks"
together with "carmody", and i did get a hit, but it wasn't what i
mean by an actual google groups usenet hit; it was a hit to some silly
"forums" archive of some sort.)
to try to figure out whether i'm making some sort of stupid mistake
(like maybe there's some sort of "don't malfunction" option that i
neglected to click on), let me ask whether other people get the same
results that i do in certain cases. for example, if i go to
<http://groups.google.com/advanced_search> and type
"y8zr5ww...@nestle.csail.mit.edu" into the "message id" input box
at the bottom and then click on "lookup message", all i get is a "your
search did not match any documents" page. does anyone know how to get
this particular search to work correctly?
--
Ralf Bader
You need not even try an actual search to see that this thing will be
defunct. The search form
http://groups.google.com/advanced_search?hl=en&q=&hl=en&
starts with "Find web pages that have...". What? Web pages?
Ralf
James Dolan
<ba...@nefkom.net> wrote:
|> to try to figure out whether i'm making some sort of stupid mistake
|> (like maybe there's some sort of "don't malfunction" option that i
|> neglected to click on), let me ask whether other people get the
|> same results that i do in certain cases. for example, if i go to
|> <http://groups.google.com/advanced_search> and type
|> "y8zr5ww...@nestle.csail.mit.edu" into the "message id" input
|> box at the bottom and then click on "lookup message", all i get is
|> a "your search did not match any documents" page. does anyone know
|> how to get this particular search to work correctly?
|
|You need not even try an actual search to see that this thing will be
|defunct. The search form
|http://groups.google.com/advanced_search?hl=en&q=&hl=en& starts with
|"Find web pages that have...". What? Web pages?
yes, it does seem pretty much insane, but i'm pretty sure that i've
sometimes nevertheless gotten acceptable results using this search
form. i don't remember exactly when was the last time that it worked
for me, but this week it doesn't seem to be working for me at all.
is there some particular other search form that i should be using
instead?
--
Bill Dubuque
> On Jul 4, 7:48 pm, jdo...@math.UUCP (James Dolan) wrote:
>>
>> my basic question is in the subject header: how badly broken is
>> google's usenet archive?
>>
>> i've often heard people complain about it, but i've usually found it
>> to be not as badly broken as the worst complaints seemed to suggest,
>> until i tried using it this week and found it to be just about
>> completely dead. is it true that it's just about completely unusable
>> now? about how long has it been that way?
>>
>> i apologize if i'm repeating questions that people are already sick of
>> answering, but because of the archive problems i haven't even managed
>> to find out to what extent the problems have already been discussed.
>
> author, was broken over the last couple of months, but I think this got
> caught up in the last week or so.
It's been broken for much longer than that - at least a few months
and it remains badly broken. I haven't seen any fixes. Indeed,
rather, it seems that it's breaking more, nor being fixed.
> Google Groups does have folks who respond to reports of broken
> functionality, though I imagine not everyone hears back about
> specific fixes.
Correction. A long time ago Google Groups had support people who
fixed bugs. The handful of problems I reported many years ago
were fixed and I actually received personal replies. Now they
no longer appear to offer an email address for bug reports and
only offer a Google groups support group, which does not appear
to be monitored by Google nowadays. I suspect the new manager is
someone too young to understand the value of newsgroups. Indeed,
many 20-somethings have never even heard of usenet newsgroups,
and, when they do, view them as antiquated compared to blogs etc.
But, of course, interesting math stays fresh for much longer periods
of time than most any other body of knowledge, so it will be a shame
if Google lets the usenet archive atrophy to the point where it can
no longer be reliably searched, etc.
Ralf Bader
I could ask the same. The search form was different until a couple of months
ago (for example, it talked about messages instead of web pages), and I
remember that precisely with the appearance of the current form things
stopped to work for me.
Ralf
Bill Dubuque
Yes, this form has been broken since at least Nov 9, 2008
(which shows you just how bad the "support" is nowadays).
The workaround is easy, e.g. see this excerpt from Slashdot [0]:
There's a bug in the advanced search form. After you do the advanced
search and it gives you the did not match any documents (e.g. [1])
just click on the "search" button on that second page. (alternately,
removing the lr=selected parameter makes it work also) (e.g. [2]).
James Dolan
bill dubuque <w...@nestle.csail.mit.edu> wrote:
|> yes, it does seem pretty much insane, but i'm pretty sure that i've
|> sometimes nevertheless gotten acceptable results using this search
|> form. i don't remember exactly when was the last time that it
|> worked for me, but this week it doesn't seem to be working for me
|> at all.
|>
|> is there some particular other search form that i should be using
|> instead?
|
|Yes, this form has been broken since at least Nov 9, 2008 (which
|shows you just how bad the "support" is nowadays). The workaround is
|easy, e.g. see this excerpt from Slashdot [0]:
|
| There's a bug in the advanced search form. After you do the
| advanced search and it gives you the did not match any documents
| (e.g. [1]) just click on the "search" button on that second
| page. (alternately, removing the lr=selected parameter makes it
| work also) (e.g. [2]).
hmm. i just tried this workaround on a couple of searches, and still
none of the searches are working for me at all.
can anyone give a specific example of a usenet search using this
search form that successfully returns a hit using the workaround, so
that i can test whether it works for me?
--
David Bernier
I tried advanced search, but I don't really like it because when it
worked, it returned discussion threads. I was looking for an
article of mine where I legally reproduced James Harris's
PrimecountH.java source code. JSH threads are often very long.
In the "Search Author's posts", there's a box that can be filled in
with, e.g.
prime -mersenne -riemann
and so on.
While searching an author's posts, only individual articles
are returned.
My post contained this:
begin PrimeCountH.java by James Harris
on one line. If I enclose it in quotes: "" and Google the Web,
I get the Math Forum article:
http://mathforum.org/kb/message.jspa?messageID=497721
On the Google the Web return of results,
at the bottom it says:
"If you like, you can repeat the search with the omitted results
included". [link in blue]
By clicking on "repeat search", there are about 300 hits,
including this:
http://groups.google.nu/group/sci.math/msg/460234dba71d6621
So I'll try the "repeat the search with the omitted results" option
some more.
David Bernier
Bill Dubuque
>>bill dubuque <w...@nestle.csail.mit.edu> wrote:
>>>
>>> yes, it does seem pretty much insane, but i'm pretty sure that i've
>>> sometimes nevertheless gotten acceptable results using this search
>>> form. i don't remember exactly when was the last time that it
>>> worked for me, but this week it doesn't seem to be working for me
>>> at all.
>>>
>>> is there some particular other search form that i should be using
>>> instead?
>>
>>Yes, this form has been broken since at least Nov 9, 2008 (which
>>shows you just how bad the "support" is nowadays). The workaround is
>>easy, e.g. see this excerpt from Slashdot [0]:
>>
>> There's a bug in the advanced search form. After you do the
>> advanced search and it gives you the did not match any documents
>> (e.g. [1]) just click on the "search" button on that second
>> page. (alternately, removing the lr=selected parameter makes it
>> work also) (e.g. [2]).
>
> hmm. i just tried this workaround on a couple of searches, and still
> none of the searches are working for me at all.
But it does in fact work on the example in [0]. Sometimes it helps
to retry a few times (server timeouts?) when Google Groups fails.
> can anyone give a specific example of a usenet search using this
> search form that successfully returns a hit using the workaround,
> so that i can test whether it works for me?
See the links in my prior post, which are from the original
Slashdot post [0] from Nov. The first fails, but the 2nd works.
Or search for "lr=" in [0] to find said message with said links.
David Bernier
I tried Bing.com with the keywords:
james dolan sci math groups google .
It's also possible to add a year, for example:
james dolan google groups 1996 ,
and I think it's good at finding archived articles.
David Bernier
Phil Carmody
They kind-of merged the web search with the news search.
There is a tick-box to actually select a 'google groups' search,
but it's not the default.
My girlfriend and I attempted to make the interface a bit more
sensible (sensible defaults, like searching usenet, for example):
http://theanna.org/goog.html
http://theanna.org/goo.html
If google's changed its interface again and those don't work please
feel free to complain, and we'll try to fix them. (Unless the search
has become completely impossible with any interface, that is.)
Phil