I've just tinkered with it, but whilst it was fast, it just
seemed...welll...kludgy. It didn't seem very flexible and just kind of
thrown together. I'm all in favor of IBM adding BIF's to do stuff, but their
XML support seems awkward - frankly, I felt like someone said "You guys need
XML support", and they dumped this in there. Those two op-codes (XML-INTO
and XML-SAX) are unlike any other opcodes/BIF's and are uuuuugly.
I admit that I haven't tried it 'seriously', but I prefer to use procedure
calls etc., which is why I like tools like RPG-XML and Scott's eXpat port
("SEP"?).
Maybe I shouldn't dis IBM - they've done their best and if they get more
peoiple using XML as a result, then great. The other thing about IBM
BIF's/opcodes is that if you need a new option, you have to petition them
and *hope* that it will make it in the next release, which could be a while
coming. With tools vendors, the time for them to provide an enhancement/PTF
is typically quicker, especially if they can enhance one bit of processing
without it requiring a test of every other bit. Mind you, with IBM you get
basically perfect software, at least as far as things like the RPG compiler
goes.
I just prefer code like this (pseudo-code):
defineHandler(type:handlerpointer);
defineHandler(type:handlerpointer);
defineHandler(type:handlerpointer);
defineParsingOptions(options);
parseXml(xml);
which is pretty much what RPG-XML does (and I think what SEP does).
Rory
On Mon, Jul 7, 2008 at 9:58 AM, Michael Ryan <michaelrtr@xxxxxxxxx> wrote:
Hi Rory -
As a matter of fact, I'm a happy RPG-XML user right now. I'm using it
for web services stuff - I hadn't thought about it for just straight
RPG-XML stuff...though that's the name...duh! :) I don't think I'm
licensed for that partition with them - I'll have to talk to them
about that.
So you find the XML* stuff to be a pain?
midrange.com
