We have a simple REST web service running on a dev server. The url is typical, e.g. http://domain/thingies/thingy5. Use it anywhere you like. We don't care. And it's really easy.
The other day I came across another team's wiki page with one of our urls in it. The page had some text discussing some code for working with the information about these thingies. This url showing up should be no surprise -- it's just the web.
But at the same time this week we're working on some other REST web services. The underlying implementation of these has to work across multiple legacy systems, several of which are accessed via their own proprietary SOAP services.
Looking at the WSDL for one, I see it's an end point definition for an interface defined in another WSDL somewhere else. But getting to that other WSDL is a royal pain -- if this is a *web* service, why can't I just *click* on something to get it?
Instead I copy and paste the pieces of text I think will form a url to the desired WSDL. Nope. Is it just gone, or did I copy and paste incorrectly?
Meanwhile next to me and in Belfast other team members are running into familiar problems with several SOAP toolkits generating code against these WSDLs. Then SoapUI makes the connection, and a co-worker cuts and pastes this into his application code instead of using one of the high and mighty SOAP toolkits. It works!
Another co-worker then suggests that maybe our app should just screen-scrape SoapUI! Brilliant! 8^)
This is the nth time in my career I've used SOAP, only to find each one a time sink more than an order of magnitude less productive than simply using HTTP. But go ahead and argue best as you can why that's a good idea.
I'll just respond, "Please don't hurt the web." Or my productivity.
No comments:
Post a Comment