[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-pim
Subject:    Re: KDE PIM Roadmap (Call to action)
From:       Guillaume Laurent <glaurent () telegraph-road ! org>
Date:       2001-06-15 21:36:16
[Download RAW message or body]

On Friday 15 June 2001 18:00, Nick Papadonis wrote:

> I could be wrong, but with large files, isn't it hard to insert or
> extract information?  Don't you have to parse from a files start for
> destination?

Yes, but let me tell you a little story. In 1997, I briefly worked for a 
company which was running a popular search engine called "Echo", which was 
indexing the french part of the Net. At that time it was indexing about a 
million pages (if I recall correctly), and served several hundred thousand 
requests a day. It had excellent rating throughout the press, and when I 
joined they had just been bought out by France Telecom (they are now 
www.voila.fr).

The machine running it was a Pentium 166, with 128Kb of memory. The engine 
was fairly simple, for each keyword you entered it would look up in its 
dictionnary and find the index to the url of pages which were related to that 
keyword.

The dictionnary was a plain flat text file.

The "dictionnary lookup" was /usr/bin/grep.

> A database would allow direct access to information.
>
> No?

If by "database" you mean things like MySQL or Postgres, then the answer is 
yes, but at the cost of overall complexity and reliability.

You really don't need a DB to manage personal info, not even if you have a 
2000 records address book (even bbdb scales to that range without problems). 
It won't buy you any performance, only problems, because it's very easy to 
corrupt a DB, rather hard to back it up, and you have to administrate it.

Something simpler like dbm might be a good answer, but I'm not sure it's even
worth the trouble.

-- 
					Guillaume.
					http://www.telegraph-road.org

[prev in list] [next in list] [prev in thread] [next in thread]