Tables Vs. XML; the data lingua franca debate.

Okay I’m exaggerating, there’s no debate, those using tables (mainly business-orientated techies/power-users) are blissfully unaware of the charms of XML; while those whose only answer to every data exchange problem is XML (or its Javascript cousin JSON) think tables are something people used to design HTML web sites before they discovered the delights of CSS – or worse, the things that DBAs (those killjoys of the development world) are always going on about.

Tables, aka datasets, aka datagrids, aka CSV or TSV files and the peoples favourite, the Excel Range, are alive and well and continue to drive most IT based business processes.

Look at this recent comparison of mashup tools over on TechCrunch, if you look at the comparison table only one product uses tables (datagrids) as its “pipe transport protocol” (Input/Output feeds) the rest are all XML variations. The product in question is Proto which is significant as it’s the only mashup tool I know (other than RSSBus which wasn’t on the Techcrunch list) that deals with the current reality of Excel/”behind-the-firewall data sources” while also looking to the near-future where more and more (but not all) data migrates to the cloud.

I not saying that XML is always the wrong choice; RSS, OPML, MicroFormats and the unloved but powerful HTML <table> tag are here to stay and more and more tools offer out-of-the-box parsing for these standards (did you know that Excel can parse a HTML table?). Likewise for formal robust interfaces between systems a bespoke XML protocol can be useful for professional developers. But if a mashup product ignores tables as an input/output format, that product is ignoring the most common business focused data format there is (and the skills of those who day-in day-out manipulate such datasets).

One other line on the TechCrunch table (see how useful tables are) that intrigued me was the assigned level of Advanced to the skill level required to use Proto. There are at least 2 million VBA users out there without any formal programming training who construct corporate internal “mashups” everyday, for that constituency Proto would be by far the easiest tool to use.


8 responses to “Tables Vs. XML; the data lingua franca debate.

  1. Tom – once again, thanks for the insightful reiteration of what we see to be a major usability issue with ‘mashup’ platforms. Namely, in our 2005 beta release of Proto, real power users were turned off because we _didn’t_ have any scripting / VBA facility. The prospect of learning a new vocabulary to do what used to be a 10 line VBA script was described as ‘working with my hands tied’.

    As such, we think think that VBA makes it easier for ‘non-programmer’ programmers (the 15 million or so power users worldwide) to get into the mix. And because these people can wrap up the ‘hard parts’ into components, they will in turn make it easier for everyone else. Thus appealing to and supporting the scripters out there (with VBA) is essential to completing the ecosystem that will support the rest of the power users.

    So by some seemingly backwards logic and some real world experience supporting the argument – we agree with you that VBA actually lowers the bar overall.

  2. Pingback: Proto « Gobán Saor

  3. Pingback: Google Spreadsheets - ETL tool « Gobán Saor

  4. Pingback: Python the new VBA ? « Gobán Saor

  5. Pingback: TAG Cubes – SQLite Star Query Part III « Gobán Saor

  6. Pingback: Excel as the iPod of Downloaded Data « Gobán Saor

  7. Pingback: SAX and Bugs and XBRooL « Gobán Saor

  8. Pingback: Excel in Data | Gobán Saor