Software development 003

Right - it's been a couple of days while I tried to sort out my next step...I've been doing a lot of reading.

I recognise a pattern here, a period of intense frustration trying to work out how to do something new without adequate documentation, and also with a great deal of scepticism about documentation. There are plenty of books &/or websites about writing HTML5 apps; but they each take a lot - a different lot - for granted. Without a deep knowledge of the field - so that you don't actually need to read the book - it's almost impossible to say whether the author is presenting some quick-and-dirty methodology for churning out rubbish so as to get their next book to press, or whether the methodology is out-of-date, or over-simplified. Very, very few writers bother to place their material in context. That annoys me.

I reckon this pattern has occurred pretty much every time I've explored some technology in the past 10 or 12 years. The kind of book I personally want isn't being published. Now, that isn't necessarily the fault of the publishers; after all, I may be an audience of one and that's not likely to be a very rewarding market. It's more likely that, actually, this situation is inevitable given the speed of development, the fragmentation of technologies, the plethora of platforms, the general overwhelmingness of it all. (How come "overwhelmingness" isn't in the dictionary? Ultimately, I've got through the frustration, worked out a tenable hypothesis for going forward, but to do it I've read (OK, skimmed) five books and about twelve websites. Plus, I've accepted that I'm going to have make the answer anyway - all that reading has got me is a decent foundation for personal research. That's this next weekend taken care of. That plus the European championships. And do you know who won the World Cup? Right, Australia, 6-1 over the Netherlands. Strangely it's gone unreported.

I enjoy this kind of thing, in a masochistic way. It's epistemologically challenging as well as being a puzzle with practical rewards for solution. Knowledge, after all, is contingent, fragmentary, constructed. The definitive encyclopaedia is a fantasy of the past. Mind you, this type of work reminds me what an attractive fantasy it is and was.

However, this blog is dual-themed; I'm trying to write an app, that's true, but I'm also thinking about other people writing apps, and in particular I'm thinking about non-specialists writing apps, be they mobile or web, as adjuncts to their non-programming job. To what extent, and how, is that possible? These are very interdependent questions. If it's very very difficult, then its being widely necessary is going to be a disaster, because it won't be possible. Maybe it can be made easy - but then (as alluded to in a previous post) we're in tool territory and that raises more layers of questions.

By analogy - Presentation managers (like Powerpoint) and spreadsheets (like Visicalc - or Excel) are tools that let people accomplish quite complex things in a more-or-less adequate-for-their-purposes kind of way. Some people, with no/little programming knowledge and less (if possible) experience write some quite startlingly complex applications in spreadsheets. Undeniably, for an IT department they're a nightmare, and when your accountant leaves and you find the entire business is documented in a miasma of incomprehensible functions, formulas and cell references you're likely to curse at least mildly, but for all that vast chunks of the business world survive on these things. And presentation managers - most presentations are terrible from an "expert" point of view, but most of them are adequate-for-purpose, and most of their creators were self-taught. Inelegant as it is - like conceding that Macdonalds is actually better than La Madeleine - good enough is actually more than good enough.

And the point of all that is - the fragmentation and dynamism of computing technology means that only extremely gifted people can be generalists covering that field. Without some intermediary tools, HTML, CSS and Javascript knowledge are of no earthly use to 99.9% of the people who aren't using them every day as core tools of their job. I can sit down and knock out a multi-user application in Excel in a couple of days, if it's reasonably simple (and it usually is). To implement the same functionality with the unholy trinity would take a month.

So, I wonder what tools there are? None that have caught the attention like the Microsoft juggernaut did in the 80's, that's for sure. And, what's the organisational cost/benefit analysis for those tools going to be?