| daniel added a comment. |
In T174028#3664170, @Marostegui wrote:
- Should we introduce content_sha1?
I don't really have an opinion on that to be honest. What would you prefer?.
I do agree that the narrower the table is, the better.
Application developers prefer to have the SHA1. I'm trying to weigh cost against benefit. I can assess benefit from the conversation on wikitech-l (basically: it would be very annoying not to have this, but not catastrophic, and maybe it's enough to have the hash in the dumps, not in the database).
Can you help me assess cost?
Also, would it help much if we used a binary representation of the hash, instead of base36?
- Should we build in support for sharing (partitioning) on the page ID?
I would prefer no partitioning for now.
I did not expect us to introduce partitioning right away. This is just about including the information needed for partitioning.
Do you also think it is not necessary/useful to provision for partion now? If we don't include the page id fields now, we would have to add them later if we want to introduce partitioning. Would that be a problem? Or would it not significantly add to the cost and complexity of introducing partitioning later?
- Should slot_revision be declared BIGINT right away?
If that maintenance is going to happen anyways, we could probably go for int with slot_revision and schedule both maintenances at the same time?
I don't know, you tell me :)
This is just my opinion, not an statement on what we should do.
It is hard to give more ideas without knowing mediawiki in depth. Actually you guys are probably more capable than me to set a real statement here! :-)
That's the problem: nobody seems for feel confident to make any final statement on questions like this. MW experts don't know the database details enough. DBAs don't know the application layer enough. Nobody is willing to do a final sign-off.
The best we can do is talk to each other and make sure neither group sees a major problem. If we wait until everyone is confident that the proposal is optimal, we'll probably never move. Apparently there just is nobody with the expertise to make that assessment.
Cc: Marostegui, cicalese, Abit, brion, Anomie, jcrespo, gerritbot, Aklapper, daniel, E1presidente, Ramsey-WMF, Jmmuguerza, SandraF_WMF, Lordiis, GoranSMilovanovic, Adik2382, Th3d3v1ls, Ramalepe, Liugev6, QZanden, Acer, Lewizho99, Maathavan, Susannaanas, Izno, Wikidata-bugs, PKM, Base, matthiasmullie, aude, Ricordisamoa, Fabrice_Florin, Raymond, Mbch331
_______________________________________________ Wikidata-bugs mailing list Wikidata-bugs@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-bugs
