After recently deciding that my own BMS (a subset of the much more powerful CMS) wasn’t up to scratch, I began thinking about the main features that a future BMS should have.

The basic starting point should be the ultimate Weblogging system. It is slightly out of date, but most of those features are the building blocks we need for a solid foundation.

There are, however, plently of features that can be added to that list to take it much further. Most are features for users, some are for developers:

• Storage: Potential for multiple storage systems (we’ll refer to them as Archives) of arbitrary type. That means the Archives can be databases, structures in the filesystem, ram drives, or any other bizarre contraption you can imagine. We want multiple storage systems so logically separate entries can be left separate. For example, the main blog entries and sidebar reading list entries are almost always unconnected.

• Extensible Text Formatter: Whatever method is used to add an entry, the processing that goes on behind the scenes to turn it into the stored format should be extensible and liberal. The bare minimum should be accepting plain text, markup, textile-formatted text, or a mixture of those. Additionally, modules could be dropped in to allow parsing of TeX, mathML or any other format.

  It would include validation processes to check that character encodings were correct, the final output was correct (i.e. it validates as the stored output format), and any other functions that the user wanted (spell checking, anyone?).

• Advanced Recall: As many ways to recall entries as possible. Current methods (permalinks and viewable archives) are fine for the moment, but throw in search and XPath queries and we start to build systems that can be viewed in ways the developers can’t imagine.

• Dynamic Categories: A sibling to Advanced Recall, Jon Udell discusses the possibilities of dynamic categories.

• Syndication: All formats should be supported natively. Let the market forces fight over which one is adopted widely while allowing users to choose the format they prefer, whether it be RSS, Atom, ESF or some other format.

• Auto-Updating: New features should be added automatically where possible. New text formatters, and formats should be downloaded and set-up transparently. End users rarely need to know how this is done.

All of this is achievable today. In the future, I may write more about how several of the features are possible. Feel free to add more suggestions (or, indeed, question the current batch) in comments.

Apparently, the web is grey. After averaging the colours of 3000 or more sites using the temporary url2png service, the output colour is #9c9c9c; a pretty bog-standard grey.

It’s not really that surprising. The majority of the web uses black text on a white background, with the various other colours pulling the colour average up towards the black end of the scale (the red, green and blue averaging out to a grey).

Still, an odd piece of trivia that we’re bound to remember.

Earlier this week, I’d been thinking about the possibilites that greatly increased bandwidth (a few Gigabytes a second) would provide; specifically, in terms of non-local computing. By utilising remote procedure call formats, it would be more than possible to do all non-trivial processing at other locations.

Today, I came across a link to the online Java compiler, via comp.lang.java.help. Although, I can’t think of any real practical use for it (other than testing if code is small and self-contained), it does show some of the potential in non-local computing.

Imagine a future version where you pass it a tar file with a huge bunch of source code (and possibly a make file), let it do the overnight processing and return a binary. That would be nice.

CVS is a powerful and useful tool. Unfortunately, running it on windows is not easy. Fortunately then, someone has produced an interesting tutorial on the matter entitled, “CVS Version Control on Windows in 10 minutes“.

I haven’t actually tried it (at the moment, I’m running CVS through Cygwin), but it seems fairly comprehensive. Definitely worth a look if you have documents that you want to keep backed-up or have more than one copy of.

Thanks to the recent IE arbitary :hover behaviour (truly a godsend to developers), cross platform CSS tabs are now easy.

The page mixes the Sliding doors technique and the new behaviour nicely. Good job, that man!