Slashdot Mirror
W3C Releases XForms
An anonymous reader submits: "On the heels of several other releases, the W3C has released XForms as a Candidate Recommendation. Coverage here and here. XForms is the way-better version of HTML forms -- it's XML-based and includes built-in client-side validation and calculations, without scripting. It is expected to replace old-fashioned HTML forms in XHTML 2.0. It's also being viewed by many as the standards-based alternative to Microsoft's XDocs. Now's your chance to try it out and submit your comments, before the official Recommendation comes out in a few months."
Very interesting, and useful too for us web developers... but its the standard besides Microsoft standard...
I thought that XForms was pretty much moribund, mostly caused (IMHO) by the "only free for non-comercial use" license. It appears that I was wrong, though. In fact, it looks like the soon to be released 1.0 will be licensed under LGPL. Too bad about the name clash, though.
Developers, please refrain from making validation so anal that it becomes unusable. People like Jennifer 8 Lee (NYT reporter) should be able to enter their name into a web form name field, for example. And sometimes people want to enter letters into phone number fields, for valid reasons (like spelling a phone number to help the recipient remember it, or to make a point). Please don't make things too restrictive.
An interesting thread to read is here in this mozilla bug
Never trust the client to do validation of data!!!! How do you think people cheat at quake etc etc... it's 'cause the final calculation is not done in the server!!!
Thank you.
-
ping -f 255.255.255.255 # if only
I strongly disagree with what you're placing in here. Client validation should be a *handrail*. It is intended to do nothing more than help the user not accidently enter a data field in a format that the server will reject, like July 30, 2002 instead of 7/30/2002. It should never fail on potentially valid data. You want heavyweight validation, use the *server*. (Matter of fact, client-side output should never be trusted, as more and more "web programmers" seem to think is a good idea).
If someone wants to enter a bogus name, they're going to do it. Overzealous "validation" is a PITA for a bunch of people. Why do it? Your "checking" above doesn't help security or the integrity of your database in the least. Now, I will grant that if someone *accidently* writes a sentence describing themselves instead of their name in the name field, then indeed, your solution would have caught something that mine would not have. However, I'm quite dubious as to the value of this (how often do people make mistakes like this?), but I'm quite certain that Jennifer and Prince are not tremendously thrilled with your refusing to accept legal names.
May we never see th
XDocs modernises the "paper shuffling" approach to business/bureaucracy, it allows one to use a "document-passing" model for business workflow, where forms are passed from person to person and a bit more information placed in the form by each person.
This is the way most government and public-sector organisations are REQUIRED to operate for auditing purposes. And it's the way most people work - the have Outlook open, they get emails with embedded Word documents, they change the document a bit, and pass it on.
XForms is still essentially just for request/response client-server programming(unfortunately).
There is NO currently standardised alternative to XDocs other than PDF forms, and PDF forms are even more obtuse to work with than XDocs. And both are single-vendor, single-implementation "standards". Currently, many companies use something even worse than XDocs or PDF Forms - they use MS Word/Excel documents with ad-hoc VB Macros. XDocs allows declarative, rather than programmatic, validation rules.
Now, unless the mainstream finally realises that they could do all this with Lisp, which is the final word in declarative/programmatic line-blurring, I predict that XDocs will be VERY popular.