On 6/7/06, Martijn van Oosterhout <kleptog@xxxxxxxxx> wrote:
Kind of all three, feeding off eachother. There's just not enough interest from any front to really get it moving. It's a fairly invasive change and without significant support and interest from somewhere, chances of completion let alone acceptance are pretty slim...
I don't get it. Maybe it's me or my environment. About half of the work we do in my company is building webs on the (excellent) eZ publish CMS. We routinely use MySQL even though the CMS (at least declaratively) wolks on top of pgsql. We use MySQL on inertia: it's the default db under eZ publish and it's easy to get hosting. On the other hand, I'd rather bite my hand off than use MySQL in some other Internet apps (e-shop with a card payment interface and others). Where am I going with all this? Well, if you don't live in an english-speaking country, there's no such thing as a single language web. Because of collation limitations, however, postgresql would be the first to be crossed out on my list no matter how good it is in all other respects. I understand that the needed change is uncomfortably invasive, but not beeing able to collate correctly is a show-stopping problem on a professional site. The user opens a drop-down with several dozens of cities, scrolls down a bit to where his city should be, dooesn't see it because the collator places the weird letter all the way down at the end of the list - and the user walks away. The other user can't get info about the bus lines to the city or he doesn't buy a product he want's because he doesn't see it where it's supposed to be. Another user notices the error and dissregards the site as amateurish. I understand I'm talking about a specific area of use, but isn't that where the biggest growth in both the number of new applications as well as the number of users is? As it is, I'm happy to have a great RDBMS to build intranet, rich client apps on - but that's about it. I've used MySQL and MSSQL (unfortunately) and they both support collation much better than pgsql. Not perfect, but much much better. Maybe postgresql should try to set less ambitious goals and instead of going for the holy grail of collation management (which is usualy praiseworthy) try to provide at least db-level collation definitions if table/row/cell level collation settings should proove too challenging at the moment. t.n.a.
