Don't Waste Time on Fedora Mailing List Archive Searches - Search Engine Broken

[David Curry]

Mark Weaver wrote:

> David Curry wrote:
>
>> Matt Florido wrote:
>>
>>> David Curry wrote:
>>>
>>>> Subject thread says it all.
>>>> The basis for my assertion is many experiences trying to find 
>>>> information in fedora archives.  Today, I made notes of steps in 
>>>> trying to find information in   fedora-config-list archives and 
>>>> conducting quite a few searches for messages dealing with pup.  I 
>>>> asked for searches on all sections, on keywords, on title, and on 
>>>> body.  I asked for searches of the entire archive and for searches 
>>>> on a couple of individual months.  In every case the result was the 
>>>> same - 0 hits.  I searched on pup + yum and again 0 hits on pup 
>>>> with a few hits on yum.
>>>
>>> [...]
>>>
>>>>
>>>> The Red Hat archive search engine is broken, broken, broken.
>>>>
>>> I've experienced something similar, but it's not broken.  Try the 
>>> extended search or just entering your search query again after it 
>>> returns the initial 0 found.
>>>
>>> Works for me.
>>>
>>> I found ~2000 hits for fedora-config-list.
>>>
>> Thanks for the feedback, Matt.  With extended search I got 25 hits, 
>> all of which were not relevant to yum or updating packages.  One was 
>> a reference to experiences "as a young pup", eniac, altair, etc., 
>> several pointed to messages concerning a book entitled "Practical 
>> Unix Programming" and the rest contained references to "pup up the 
>> volume".  Entering the search query again after an initial return of 
>> 0 found, returned 0 -- again, and again, and again.
>>
>> If Fedora archive searches work for you (sometimes) and for Rahul, 
>> but not for me then they are not completely broken -- just random, 
>> blatantly unreliable events.
>
>
> but doesn't the mere fact that you had to go through specific "extra" 
> steps to work around the primary search function mean that it is 
> indeed broken on some level? broken == doesn't work according to 
> initial design.
>
No quarrel here, Mark!  Even the specific "extra" steps were total time 
wasters for me. 
Referring anyone directly to the list archive instead of to the archive 
through an alternative search engine (or engines) such as Google is 
likely a disservice.