On 26/08/2024 20:06, Juan C. Sanz wrote:
TLTR. Propose Take Firebird out of experimental mode and explain
author’s opinion about current situation of Firebird implementation, and
the benefits of put it in service.
Background
Base is undoubtedly the ugly duckling of LibreOffice.
The reasons why, in my opinion Base is not more used are
*
The only type of embedded databases that the normal user sees when
he wants to create a database is HSQLDB. Since the days of
OpenOffice, these databases have a reputation for being slow, and,
more importantly, for corrupting data, rendering the database unusable.
Imho the reason is actually much more fundamental. If you haven't been
to Uni and done a CS degree, you haven't (a) been indoctrinated into the
Relational religion, and (b) you don't understand all this normal form
stuff.
(Note that most CS grads seem to know only the relational model. They
don't know about any other databases. It's like thinking studying the
Tudors makes you an expert on British History ...)
*
Databases are less and less used in small businesses and households,
having been replaced by “databases” made in spreadsheets.
I work for a BIG business (FTSE 100 (or was, I think we fell out a
couple of quarters ago)), and we have a massive database implemented as
loads of Excel sheets. The problem is it "just growed", because we are
Business Analysts, not Computer Analysts (well, I consider myself a
Database Analyst, but that's from a previous life, not my current role).
*
Databases appear to be very difficult to understand for a normal user.
Imho, it's RELATIONAL databases that normal users can't understand ...
*
User interface is somewhat outdated
*
Base does not provide some facilities that other more popular DBMSs do.
Make that *R*DBMSs.
Firebird implementation. Back to experimental mode
I'm not going to go on much more, I may be a fanatic but I don't want to
put everybody off :-) But I'm very much a fan of NOSQL databases, in
particular MultiValue. (NoSQL comes in three versions according to
Wikipedia - v1 was a database, v2 is "Not only SQL" ie key-value plus
schema such as MultiValue, and then v3 is pure key-value like Amazon's
DynamoDB and there's a Google version, all without schema and aimed at web.)
If you want to play with a MultiValue database, play with ScarletDME
(https://github.com/WolWise1/ScarletDME/tree/dev/). It was originally
written as an embedded database, so it could be embedded into
LibreOffice. And then you could also install the whole shebang, and use
the database tools as well. (Note that it's currently linux only,
unfortunately.)
At work I'm planning to replace all these Excel spreadsheets with an
instance of OpenQM (the commercial version of ScarletDME). And given
that most historic implementations of MultiValue databases were built by
*users*, not CS guys, surely they have to be easy for your typical user,
no? I'm hoping (and expecting!) my fellow analysts to adopt it very
rapidly once they realise the power and ease of use of the MultiValue model.
Feel free to email me privately (that's anyone!), who wants to discuss
it. I'm happy to discuss things on-list, but I don't want my enthusiasm
to alienate people!
If I start working on OpenQM at work, hopefully that will also give me
the drive to work on ScarletDME privately ...
Cheers,
Wol
