Re: EU.net and eu.net database entries
Marten,
Sorry Piet, old, old, old software ... The new software will take care of this, but I think the new software will change everything to lowercase. I have not written this part yet, have to put some thought in it though.
Anyone in the database working group has any feeling how these should be dealt with ? From a programming point of view, handling everything lowercase is much easier, I do not really want to check every single combination of lower/uppercase before I update something ... Then again, after some thinking (I usually type faster than I think ;-) the indexing is done in lowercase, and key lookups are always normalized to lowercase before a lookup.
So the real question is, should entries in the database be left untouched when it comes to upper/lowercase, or should all be normalized ?
I'd prefer the following behaviour: - In the external representation keep it exactly as it was submitted, with mixed case intact. - Internally, whenever you use this string for an index, compare or lookup, consistently convert it to lowercase, thus prohibiting duplicates with different case. I think this would even be in line with the current behaviour of other programs... I don't have any idea what this would mean to the implementer programming-wise... Wilfried.
"Wilfried Woeber, UniVie/ACOnet" <woeber@cc.univie.ac.at> writes * Marten, * * >Sorry Piet, old, old, old software ... The new software will take care of * >this, but I think the new software will change everything to lowercase. I * >have not written this part yet, have to put some thought in it though. * > * >Anyone in the database working group has any feeling how these should be * >dealt with ? From a programming point of view, handling everything lowercas * e * >is much easier, I do not really want to check every single combination of * >lower/uppercase before I update something ... Then again, after some thinki * ng * >(I usually type faster than I think ;-) the indexing is done in lowercase, * >and key lookups are always normalized to lowercase before a lookup. * > * >So the real question is, should entries in the database be left untouched * >when it comes to upper/lowercase, or should all be normalized ? * * I'd prefer the following behaviour: * * - In the external representation keep it exactly as it was submitted, with * mixed case intact. * * - Internally, whenever you use this string for an index, compare or lookup, * consistently convert it to lowercase, thus prohibiting duplicates with * different case. I think this would even be in line with the current * behaviour of other programs... * * I don't have any idea what this would mean to the implementer * programming-wise... This sounds good. In fact, looking through the code as it is now, this is sort of the bahaviour I am aiming at. A little more on the software. Now that things have cooled down a bit (no more IETF and stuff, and vacation season) I am fulltime on the new software. Currently I (and Tony) are working on the last but most difficult pieces of the software, the update mechanism and syntax checking. We are being fairly strict in terms of syntax errors. As soon as I have finished the syntax part of the code, I will summarize on this list to indicate what syntax is acceptable and what will generate errors and or warnings. My guess now is that in a week or 2, we would like some people to give the software some beta tests. We would like a small number of people who have some time to put in this, so that we can shake out the most obvious bugs asap. -Marten PS Of course all the documentation needs to be written ;-)
participants (2)
-
Marten Terpstra
-
Wilfried Woeber, UniVie/ACOnet