1. RE: Searching the Mailing Archives

bensler at mail.com wrote:
> Everytime I delve into the mail archives, I quickly became frustrated.
> 
> I do a search, I get the first 50 results (sorted by relevance?)
> How do I get the rest of the results?
> 

Yeah, I was thinking someone ought to make a proper search facility for 
the archives.  Would it be possible to get ahold of the raw archive 
files?

new topic     » topic index » view message » categorize

2. RE: Searching the Mailing Archives

Robert Craig wrote:
> Chris Bensler writes:
> > Everytime I delve into the mail archives, I quickly became frustrated.
> > I do a search, I get the first 50 results (sorted by relevance?)
> 
> Yes. They're sorted by score.
> The scoring system tends to rate messages that match multiple
> keywords somewhat higher than messages that match
> the same keyword many times.
> 
> > How do I get the rest of the results?
> 
> You can't, but you can try narrower ranges of dates,
> specify sender, extra keywords etc.

I was thinking something along the lines of compiling a list of posts 
that contain useful code snippets. (Just an example of numerous 
scenarios where I would have liked to see all the results)

Search for '=' and '+' and 'end' to filter out msgs that don't contain
code.
I SHOULD have a list of hundreds of posts.

If I want ALL of the relevant posts, I have to search every range of 
dates, being sure that I have less than 50 results for each range 
(otherwise I may have missed a post)

> 
> > I do a search for "pass by reference", and I get results 
> > for every post that contains EITHER 'pass', 'by', or 'reference'. 
> > How do I specify that my filters are inclusive, not exclusive?
> 
> Phrases inside quotes are not supported.
> You should search for:  pass by reference
> without quotes.
I used the quotes in my post to segregate my filter strings. I didn't 
actually use the quotes when I did the search.

> 
> > How do I sort by Author? by Date?
> 
> You can't.
> 
> > Is it possible to have the MIME stripped from the posts 
> > that are in the archives?
> 
> To conserve space (there's over 50Mb of message data),
> most MIME attachments are stripped out before the message
> is saved. In some cases people actually put the body of their
> message in MIME + HTML, so if there's any doubt, the MIME is retained.
> 


> The search program is written in C.
> Euphoria for FreeBSD is now very stable.
> I'm tempted to rewrite the search in Euphoria
> just to show that it won't be much slower.
> Euphoria code is (of course) much easier to
> work with than C, so a few extra features might get added
> along the way. No commitment - I really should work on other things.
> 
> Regards,
>    Rob Craig
>    Rapid Deployment Software
>    http://www.RapidEuphoria.com

Is this something that a user could write?
EU CGI in linux is possible, and from what you have said, EU for FreeBSD 
almost perfectly maps to EU for Linux. If there is anyone out there who 
is willing to do it...

If I knew CGI, I would offer my aide. *shrug*

Chris

new topic     » goto parent     » topic index » view message » categorize

3. RE: Searching the Mailing Archives

How about searching by thread?

If the search engine were to be complete, it should support searching by 
thread.

Chris

bensler at mail.com wrote:
> Robert Craig wrote:
> > Chris Bensler writes:
> > > Everytime I delve into the mail archives, I quickly became frustrated.
> > > I do a search, I get the first 50 results (sorted by relevance?)
> > 
> > Yes. They're sorted by score.
> > The scoring system tends to rate messages that match multiple
> > keywords somewhat higher than messages that match
> > the same keyword many times.
> > 
> > > How do I get the rest of the results?
> > 
> > You can't, but you can try narrower ranges of dates,
> > specify sender, extra keywords etc.
> 
> I was thinking something along the lines of compiling a list of posts 
> that contain useful code snippets. (Just an example of numerous 
> scenarios where I would have liked to see all the results)
> 
> Search for '=' and '+' and 'end' to filter out msgs that don't contain
> code.
> I SHOULD have a list of hundreds of posts.
> 
> If I want ALL of the relevant posts, I have to search every range of 
> dates, being sure that I have less than 50 results for each range 
> (otherwise I may have missed a post)
> 
> > 
> > > I do a search for "pass by reference", and I get results 
> > > for every post that contains EITHER 'pass', 'by', or 'reference'. 
> > > How do I specify that my filters are inclusive, not exclusive?
> > 
> > Phrases inside quotes are not supported.
> > You should search for:  pass by reference
> > without quotes.
> I used the quotes in my post to segregate my filter strings. I didn't 
> actually use the quotes when I did the search.
> 
> > 
> > > How do I sort by Author? by Date?
> > 
> > You can't.
> > 
> > > Is it possible to have the MIME stripped from the posts 
> > > that are in the archives?
> > 
> > To conserve space (there's over 50Mb of message data),
> > most MIME attachments are stripped out before the message
> > is saved. In some cases people actually put the body of their
> > message in MIME + HTML, so if there's any doubt, the MIME is retained.
> > 
> 
> > The search program is written in C.
> > Euphoria for FreeBSD is now very stable.
> > I'm tempted to rewrite the search in Euphoria
> > just to show that it won't be much slower.
> > Euphoria code is (of course) much easier to
> > work with than C, so a few extra features might get added
> > along the way. No commitment - I really should work on other things.
> > 
> > Regards,
> >    Rob Craig
> >    Rapid Deployment Software
> >    http://www.RapidEuphoria.com
> 
> Is this something that a user could write?
> EU CGI in linux is possible, and from what you have said, EU for FreeBSD 
> 
> almost perfectly maps to EU for Linux. If there is anyone out there who 
> is willing to do it...
> 
> If I knew CGI, I would offer my aide. *shrug*
> 
> Chris
> 
>

new topic     » goto parent     » topic index » view message » categorize

4. RE: Searching the Mailing Archives

> Is this something that a user could write?
> EU CGI in linux is possible, and from what you have said, EU for FreeBSD 
> 
> almost perfectly maps to EU for Linux. If there is anyone out there who 
> is willing to do it...
> 
> If I knew CGI, I would offer my aide. *shrug*
> 

I could do it if I had the archives to work with, but it would be in 
PHP, not Euphoria.  Which would probably mean putting it up on a new 
site somewhere, but that's not a particular problem either...

new topic     » goto parent     » topic index » view message » categorize

Search



Quick Links

User menu

Not signed in.

Misc Menu