[Monetdb-developers] The new "ws-API", Algebra & XRpc (& PFtijah)

Stefan Manegold Stefan.Manegold at cwi.nl
Fri Oct 6 15:08:00 CEST 2006


On Fri, Oct 06, 2006 at 02:39:19PM +0200, Jan Rittinger wrote:
> If this wsid is a concept that is strictly necessary for the update 
> support (and nobody except me thinks that this bit shifting might be not 
> the nicest solution) we could also think about getting rid of the ws and 
> replacing it with the wsid. I just do not like the fact that there are 
> two variables underway that have some side effects on each other.

To be honest, I'm still in the process of analyzing trying to understand,
what all the flavors, pros, cons, and ideas behind having/using wsid are;
hence, I cannot tell, whether it is "strictly necessary" or not.

As far as I understand so far, the basic idea behind the "bit shifting" is
merely to store both the id of (and hence reference to) BAT ws as well as a
unique identifier --- that is still valid after a query/transaction (and hence
cannot simply be the ws BAT id, which might be re-used for a subsequent
query/transaction) --- in one single atomic value.

Further, I agree, that having two related variables that need to be kept in
sync and treated accordingly is not very handy for maintenance --- my guess
is, that the current situation resulted from lack of time and fear to
implement (something like) your (d) proposal; which I agree is "nicer" &
"cleaner", but requires quite a lot of possibly error-prone code changes ---
nevertheless, I might give it a try during the weekend ...

Stefan

> So a fourth proposal (d) would be to replace the occurrences of ws by 
> wsid. This would mean that *all* document accesses would have to include 
> another indirection (even all path steps).

-- 
| Dr. Stefan Manegold | mailto:Stefan.Manegold at cwi.nl |
| CWI,  P.O.Box 94079 | http://www.cwi.nl/~manegold/  |
| 1090 GB Amsterdam   | Tel.: +31 (20) 592-4212       |
| The Netherlands     | Fax : +31 (20) 592-4312       |




More information about the developers-list mailing list