From: bap@shrdlu.com   
      
   On 2026-01-18, Someone Else  wrote:   
   >   
   > It may be possible, now, to sort the way you want. But there are some   
   > problems to be solved, first. Among them are:   
   >   
   > * Defining the proper sort. Do you sort Mac & Mc together? Do you   
   > ignore leading articles? Even French ones (such as Les)? Let's not   
   > forget Gary McGath's complaint: 'I have a song in Fourscore called "A   
   > Is A." The app alphabetizes it under "I," treating "A" as a definite   
   > article.' How about the apostrophe in O'Brien? And remember that your   
   > sort order has to be defined to take into considerations you've never   
   > come across.   
      
   This is a problem that was solved, for English, in the 19th century when   
   librarians needed to decide the right order for shelving books.  I was a   
   member of the Library Association's Cataloguing and Indexing special   
   interest group.We didn't ever have to revisit this issue.  It's a done deal   
   and has been for more than a century.   
      
   Modern database software lets you select the collating sequence for a table   
   based on each languages' rules.  That becomes the default sort order for   
   reports.  But you can change that programmatically.  Reports should always   
   use a collating sequence that puts the data where the reader expects it to   
   be.  The physical format of the data on disk should not make any difference.   
      
      
      
   >   
   > * Do you want the same sort for everything, including book titles,   
   > and proper names? How do you reasonably sort people's names if you   
   > don't know whether the surname precedes or follows the given name?   
   >   
   > * You need to convince enough people that your sort order is not only   
   > more nearly correct than anyone else's, but that it's worthwhile   
   > changing to it.   
   >   
   > * Figure how to code the sort in every programming language in use,   
   > for all of the Unicode character sets. Remember that they can be   
   > combined. Also, there are many Unicode sets not yet defined, and the   
   > algorithm must support those new ones.   
   >   
   > It's correct to both hope and expect that technology will keep   
   > improving in ways that better help us. I think it's naive to expect   
   > that it'll ever be perfect. It's even more naive to believe you can   
   > persuade everyone that there is a One Best Way (TM) to do things.   
      
   Or that once you have documented it as an international standard that they   
   will stick to it, or even check whether a standard already exists before   
   they reinvent the wheel.   
      
   --   
   Bernard Peek   
   bap@shrdlu.com   
   Wigan   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)