The Lack of a Small Unified Database
The database interface in KOffice is called Kexi. Kexi is an integrated environment for creating database schemas and for inserting, querying and processing data. It can run without KDE, on UNIX, MS Windows and Mac OS X. Kexi already creates self-contained database files (.kexi). Queries and other metadata are stored inside the database itself, in special hidden tables named kexi__*. Such metadata is of both visual (column widths, detailed cell formatting) and functional natures (constraints, error messages). This is quite different from what happens today in OO.o, where metadata is stored in an XML package while the data itself is stored somewhere else.
Obviously, using the Microsoft Access format, .mdb, is not an option. Why remain dependent on yet another proprietary format that could change overnight? As far as I know, the best solution is SQLite, a lean and mean database engine already available in KDE and GNOME--it's embedded in Kexi and gnome-db, respectively--and built from less than 30K lines of C code. Using SQLite, databases can be designed and used with a standalone, public domain browser on Linux, Windows and Mac OS X. Wrappers in all popular languages already exist; even PHP 5.0.0 already embeds SQLite, version 2.8.14.
Another reason to go for this engine might be the SQLite license, making it suitable for inclusion in proprietary products. The main drawback of SQLite today is the lack of some functions, such as ALTER TABLE, check constraints and referential integrity. However, the addition of these function, starting with ALTER TABLE, is planned to happen over the next months. Last but not least, the small amount of code involved makes it easy to become familiar with SQLite and add new features. D. Richard Hipp, the author of SQLite, is very interested in seeing this project succeed.
Both developers and end users can contribute to make this portable database dream come true. The first thing to do is help embed SQLite in OO.o. A possible approach to this goal is discussed here. Other things you should know are the "Basic UNO" and "Database Access" sections of the Developer's Guide, as well as the relevant APIs of SQLite. The list to join is email@example.com. The developers eagerly are awaiting your support on this particular sub-project.
The hardest part, of course, is to make a complete standard of all this. Once OO.o embeds SQLite, what still would be missing from making it possible to exchange databases and forms directly among the users of OO.o, KOffice and, later, everybody else? Above all, do the KOffice and OO.o teams have the will to converge fully on achieving this standard?
After hearing my arguments, OO.o developers agreed that SQLite indeed might be a better bet than HSQLDB, even if it will require additional time to implement. Sun already has a strong interest in standards, so I hope it will support this proposal of mine. As far as KOffice is concerned, all its core developers have the goal that "all KOffice applications either follow the OASIS format or help to define new OASIS definitions for formats that are currently not standardized". Looking good, no? Of course, SQLite is only the database engine part. The greatest work is to define and use in the same way all the other associated information--how reports and forms are represented and recorded, how the query is stored, and so on. Theoretically, it should be possible to converge on some XML user interface description--perhaps UIML, another OASIS standard.
Another issue that needs to be dealt with is the current difference in file format approaches. In Kexi, everything is added to the actual database in separate tables. In OO.o, a ZIP archive with administrative XML streams is used. As of current plans, OO.o 2.0 is supposed to embed the actual data into that archive. Also, OO.o forms and reports are real OO.o Writer documents generated by a wizard that contain a macro that, when the document is opened, fills it with the data. Consequently, either Kexi or OO.o should change the top-level file format itself (ZIP archive or SQLite DB) and then agree on a common formats for queries and forms. Initially, the choice seemed to me to be already made, even if not all the interested parties had realized it yet. If the OO.o format has become an open OASIS standard, and KOffice is adopting it as its own native format and Kexi is part of KOffice, then what's left? The standard forms will be OASIS text documents, as they already are in OO.o. But, this is my initial opinion. What matters is that one common format does comes out and that it is given the rank of official standard. To me, SQLite seems to be the best way to do it.
Articles about Digital Rights and more at http://stop.zona-m.net CV, talks and bio at http://mfioretti.com
- My Childhood in a Cigar Box
- Papa's Got a Brand New NAS
- Applied Expert Systems, Inc.'s CleverView for TCP/IP on Linux
- Tech Tip: Really Simple HTTP Server with Python
- Let's Go to Mars with Martian Lander
- Rogue Wave Software's TotalView for HPC and CodeDynamics
- Panther MPC, Inc.'s Panther Alpha
- Simplenote, Simply Awesome!
- NethServer: Linux without All That Linux Stuff