Lists Home |
Date Index |
Where once we would design EDI (EDIFACT) messages, now we design XML messages. XML is just fine for representing the kind of hierarchical structures that EDI uses. I don't know that it does a better job of it than EDI, but it has the advantage of being ubiquitous.
On the other hand, XML actually can have a big disadvantage over EDI, the same one that is its main advantage, actually. Not many people ever thought of using EDI everywhere, for for anything; it is okay over the wire, but it is normally transformed into something more malleable as quickly as possible, once received. In theory XML is a step up from EDI, because there is a whole raft of tools available to help you to transform it.
Unfortunately, some organisations appear to be taking the position that because XML is now usable in every tool under the sun, that not only should it be used everywhere, but it can be used everywhere as is. So, at least where I work, I get the distinct impression (hopefully false) that organisations are actually wanting the data interchange format to be what they build new back end systems over, so they don't have to do bother with any transformation. This seems an amazingly short-sighted, and dangerous thing to do. A technology touted as an aid to loosely coupling disparate applications, is, ironically, leading to tighter coupling than existed before.
I think this problem is exasperated by organisations such as the one I work for. If you sign up to use a standard within your vertical industry, and send people along to committees to influence message design, you get a false sense of being in control. This will presumably evaporate once member organisations begin exchanging data with organisations outwith the standards body, who refuse to use our standards. That'll learn 'em.
>Paul Prescod wrote:
>XML is weird for business data? Did you ever work with EDI?
EDI isn't weird, it is actually very simple, it just looks terribly complicated. For a company wanting to sell EDI based software this is a godsend. The software is fairly trivial to put together, but because EDI looks hard to your average consumer, it is quite easy to convince them to part with lots of money, firstly to use the software, and secondly to have someone else set it up and maintain it for them. This gives the software vendor a nice, steady stream of recurring revenue for hardly any work.
XML has suffered from the problem of looking too simple to the user. Whilst this has helped uptake, users of XML expect to get it for free, or less. Fortunately a lot of people are putting a lot of effort into making XML seem as hard as EDI, and I think their efforts are beginning to pay off.