Re: One True i18n Table
Date: Thu, 14 Oct 2004 10:34:08 -0400
Message-ID: <1p2mkc.e48.ln_at_mercury.downsfam.net>
Laconic2 wrote:
>
> "Kenneth Downs" <firstinit.lastname_at_lastnameplusfam.net> wrote in message
> news:9nolkc.757.ln_at_mercury.downsfam.net...
>> Like many of us, at one time I "invented" the one true lookup table, >> eventually saw the error of my ways, and stopped doing stuff like that.
>
> I little while ago, I described how a created a view with 185 selects and
> 184 UNIONs in it so that I could have the appearance of OTLT without the
> materialization of it.
>
> "Skunk works"
>
> Did you see that discussion?
I did. Definitely a case for a code generator!
I came within an inch one time of doing the opposite, using the ELEMENT_TYPE column in a OTLT to generate any number of views, making it look like there were so many. By way of disclaimer, I did not create the original spec for that situation, and was dealing with somebody else's OTLT. They were convinced they had invented the idea, aren't we all?
>> >> With one exception. I find I still use a single lookup table to hold all >> messages for internationalization. This table includes descriptions of >> objects, error messages, notifications, and so forth. Another table >> holds languages, and a third cross-references them. >> >> Oddly enough, the reasoning is the exact same as for the OTLT, it is in
> fact
>> easier to pull those descriptions out of a single place than to have a >> sub-table for every table in the system. >>
>
> My question is this: do you experience the same downside that you did
> with
> OTLT, or don't you?
As I said, I have no difficulties with it. In examining the situation, the only factor I can identify is that user's don't do the input, so I don't have to worry about validation. Am wondering if others have had the same experience.
-- Kenneth Downs Use first initial plus last name at last name plus literal "fam.net" to email meReceived on Thu Oct 14 2004 - 16:34:08 CEST