[Kde-pim] Akonadi: single database design mistake?

[bhlevca]

I don't want to waste your time replying to worthless emails, but some things
need clarification.


Milian Wolff wrote
> Considering that I did a bunch of profile runs on kdepim, I rarely, if
> ever, 
> saw a bottleneck in the database. Mysqld is also running just fine.
> 
> What we have and what we are working on, is an imported virtuoso/nepomuk 
> integration as that is quite often slow. But its getting better and
> better. 
> So: stay tuned and "calm down" as you also say :)
> 
> .... 
> 
> Yes, but that has nothing to do with the "single database" that akonadi
> uses. 
> We actually use two ;-) And the virtuoso/nepomuk one is the slow(er) one,
> not 
> mysql.

What I mean by database centric is not the mysql database, but the
akonadi/nepomuk/virtuoso handling everything. 
As you pointed out mysql is actually doing fine. That is just a backend
which can be switched to sqlite or postgres.


> Ah, what is constructive and what not lies in the eyes of the beholder.
> While 
> you used more words, its essentially the same as "meh, it doesn't work for
> me, 
> but look - XYZ works for me". While its an interesting hint it won't help
> much 
> if at all. And pointing at the single database is just wrong as that's
> _not_ 
> the problem.

As you used more words to defend the current  _not_woking_good_enough_
design.  It is not a "meh" as I spent enough time to try to make it work
decently and I was not throwing flames I was just trying to suggest that it
is a good time to look back and decide if the current "all integrated" is a
viable solution before everyone abandons it.

I understand that is is your baby and as any parent it is hard to find and
faults in it, but you should at least listen, because I and the people
before me didn't mean to put you down but raise issues. The 'meh' does not
work for me is not a nice thing to say. 


> If someone would write such a second backend and would maintain it... But 
> noone is going to do that. Or will you write a backend for notmuch?

That's why you should keep things simple. If you has the men power Oracle
has this design would probably work. I would write a search engine based on
notmuch approach, but not in "the current all integrated framework". It
would have probably a better speed but not by much.  

I guess that we won't agree on this one so let it rest. I just wanted to
suggest that maybe it is time to reevaluate some design decisions and their
feasibility.
 
Cheers and good luck,
Bogdan




--
View this message in context: http://kde.6490.n7.nabble.com/Akonadi-single-database-design-mistake-tp1053901p1528667.html
Sent from the kde-pim-general mailing list archive at Nabble.com.
_______________________________________________
KDE PIM mailing list kde-pim at kde.org
https://mail.kde.org/mailman/listinfo/kde-pim
KDE PIM home page at http://pim.kde.org/