Re: thank you IBM
Andrew Hamm and Serge Rielau (IBM DB2 SQL compiler) debated ANSI SQL compliance. Hamm called the standard over-analysed and inflexible; Rielau said non-standard features mean lock-in and that several groups reviewing a requirement yields better abstraction than pragmatism. Hamm said he could no longer cite examples from the old Informix manual. No technical resolution; an opinion exchange.
Auto-generated by Claude from the posts below — may be imperfect; read the full thread.
Serge Rielau wrote: > population of non-ANSI DBs. Some of those non-ANSI properties are > "merely" extentions and can be made to fit into DB2, others can simply > not be married with ANSI. > Quite a conundrum. So give us a special type of database, not unlike informix tends to treat ANSI as special type...
Andrew Hamm wrote: > Serge Rielau wrote: > >>population of non-ANSI DBs. Some of those non-ANSI properties are >>"merely" extentions and can be made to fit into DB2, others can simply >>not be married with ANSI. >>Quite a conundrum. > So give us a special type of database, not unlike informix tends to treat > ANSI as special type... Personally I have moved past the denial phase. Trying hard to cross from anger to acceptance of the inevitable. Cheers Serge -- Serge Rielau DB2 SQL Compiler Development IBM Toronto Lab
Serge Rielau wrote: > Personally I have moved past the denial phase. Trying hard to cross > from anger to acceptance of the inevitable. You are a fan of ANSI? You think that non-ansi databases are wrong? Personally I think ANSI standard has, as usual, been a victim of over-analysis by a self-important committee who don't necessarily have a lot of practical experience or ability to think of better ways. So many of their assumptions are too inflexible and clumsy.
Andrew Hamm wrote: > Serge Rielau wrote: > >>Personally I have moved past the denial phase. Trying hard to cross >>from anger to acceptance of the inevitable. > > > You are a fan of ANSI? You think that non-ansi databases are wrong? They are non-standard. Being non-standard means lock in. That's all. > Personally I think ANSI standard has, as usual, been a victim of > over-analysis by a self-important committee who don't necessarily have a lot > of practical experience or ability to think of better ways. So many of their > assumptions are too inflexible and clumsy. That may be true in some cases (got examples?). However, in general having different groups of people look at the same requirement leads to a level of abstraction a single group of developers can't achieve. Pragmatism wins too often, orthogonality goes down the drain. I live is a constant love-hate relationship with my friends from the other 2 same DB2 codebases. They always propose everything wrong.. and they always think what I propose is wrong ;-) Just because soemthing requires only 10 lines of code changes in DB2 UDB for LUW or Informix IDS, doesn't make it right for the language. Cheers Serge -- Serge Rielau DB2 SQL Compiler Development IBM Toronto Lab
Serge Rielau wrote: > That may be true in some cases (got examples?). > However, in general having different groups of > people look at the same requirement leads to a level of abstraction > a single group of developers can't achieve. Many years ago I have fond (no, not) memories of reading the ANSI database details in the Informix manual, and repeatedly shaking my head at the pointless OCD assumptions. Have examples? Not any more. Am I going to re-read the manual just to make a list of silly assumptions? no.