Home > Drizzle > Unicode schemas - IMHO, a horrible idea

Unicode schemas - IMHO, a horrible idea

December 7th, 2008 Leave a comment Go to comments

Stewart, Brian and Jay were having some fun and were able to implement support for Unicode characters in Drizzle schemas.

I pretty sure that this is not allowed in -any- massively used database. So +1 for Drizzle right?

In my mind, not so much.

First off, it should go to show that if the main website linking to your blog can't even display the character correctly, maybe it's time to rethink if it is something worth keeping.

Second, while this helps prove that everything in Drizzle will be Unicode capable, it creates an utter nightmare for supporting schemas created with high ascii or foreign characters. Some of us are already supporting other people's existing Latin-based schemas. Now add in the need to have a character map open and terminal windows that can copy/paste/display Unicode characters, and the difficulty of working with these has increased ten-fold. Not to mention the filenames storing the databases/tables on the filesystem will become harder to work with too.

Lastly, look at everything out there that -is- Latin-based and nobody is affected, for example:

  • The programming languages themselves - C, C++ (PHP, Java, whatever)
  • Existing schemas in any DB
  • Domain names (Unicode support doesn't seem to have taken off still)

It is just a matter of life that native Chinese people have to develop code in Latin-based characters, have to design their schemas (currently), and have to visit the majority of websites (if not all) in Latin-based characters. I don't see anyone trying to change C or C++ to support Unicode characters... there just isn't a real reason for it.

To summarize, in my opinion - what seems cool today will become a headache in the future if unleashed. I am totally in support for end-to-end (and only) Unicode support for data and such, but allowing it in schemas are a bit too much.

Categories: Drizzle