- From: T. V. Raman <tvraman@us.ibm.com>
- Date: Fri, 11 Mar 2005 16:49:18 -0800
- To: jeacott@hardlight.com.au
- Cc: tvraman@us.ibm.com, www-forms@w3.org
Ah but that's the question -- 99% of which tasks? the tasks you knew about yesterday? The ones you knew about last year? The ones you dont know about yet, but that we might collectively think of next year? As I tried to outline, XForms defines constructs that people have by repeated scripting experiments proved are needed and possible in the browser. Put another way, to invent the notation for a double integral, A) you need to discover the concept of integration; B) You need to invent a notation for it-- C) Reason with that notation until you become adept with the concept which is initially new and requires many paragraphs to explain; D) Having seen the light, get the AHA moment where you realize that you can apply the concept repeatedly to integrate over multiple dimensions. I for one am not going to take on the task of inventing notations for things not yet known; to steal a line out of Sir Humphrey's book in yes Minister, "I foresee all kinds of unforeseen problems there". >>>>> "Jason" == Jason Eacott <jeacott@hardlight.com.au> writes: Jason> I want to see a spec for X-forms that's complete Jason> enough that 99% of tasks can be performed WITHOUT Jason> script, since it seems unlikely that any Xforms engine Jason> in the future will possibly be able to make sense of Jason> all the script that could end up as part of an Jason> x-form. by that I mean that a browser based Jason> implementation of xforms is unlikely to read and Jason> understand script targeted at an open office document Jason> for example, or some other-modal variety of renderer. Jason> if this isnt the case then the second you use script Jason> you are no longer x-platform compatible. even relying Jason> on css (which lots of xforms I have seen seem to do) Jason> will only work in css capable browsers, so the Jason> embedding of an Xform inside something like Xhtml will Jason> only get you as far as Xhtml compatibility at best. so Jason> If you want a truly x-platform x-modal xform (this Jason> gets a bit weird since i understand here that the Jason> x-form itself is wrapped up in other markup) you need Jason> to embed it in some non-specific xml wrapper and then Jason> alter the wrapper depending on target platform(with Jason> xslt or something presumably). wouldn't it have been Jason> better to just bundle up an xform as a discrete unit Jason> the could reliably be sent to any device, independent Jason> of the wrapper you choose to surround it in? after i Jason> don't think there is any guarantee that just because Jason> some device can render an x-form that it can also read Jason> xhtml? so there is always a chance that the behavior Jason> of such a device could be to just ignore any markup it Jason> didn't understand, including anything inside it. this Jason> is pure specultaion but I hope you see where I'm going Jason> with it. Jason> Jason> Cheers Jason. Jason> Jason> Jason> Jason> Jason> Date sent: Fri, 11 Mar 2005 14:59:26 -0800 To: Jason> andyh@agaricus.co.uk Copies to: www-forms@w3.org Send Jason> reply to: tvraman@almaden.ibm.com From: "T. V. Raman" Jason> <tvraman@us.ibm.com> Subject: Re: XForms Myths Exposed Jason> - By Ian Hixie (Opera) Forwarded by: www-forms@w3.org Jason> Date forwarded: Fri, 11 Mar 2005 22:59:47 +0000 Jason> >> > > As a follow-up to all that has been said here: >> >> >> The Web Forms fans' attempt to polarize between XForms and >> scripting is unfortunate --- and based on what I've heard >> them say about XForms, I'm not willing to judge whether >> that is a consequence of ill-will, a specific intent to >> drive a personal agenda, or simply miscommunication. >> >> The XForms designers ---myself included have always >> motivated the XForms design by pointing out how things >> that are complex to do on the legacy HTML4 scripting Web >> become dramatically easier on the XForms + XHTML >> Web. However, in expressing this, and in our focus on >> XForms, we may not have communicated our position on the >> role of scripting very well. >> >> I believe that scripting played a key role on the Web >> during the late 90's in helping Web authors discover the >> next set of things beyond HTML4 that would like. Since the >> script interpreter was built into the browser, advanced >> Web programmers could experiment to their heart's' >> content, and what's more, even deploy the result of their >> experiments to a wide audience to do real-live usability >> testing. I believe this to be a first in the somewhat >> fledgling field of user interface engineering. Now that >> we have done this level of experimentation, it is time for >> the next turn of the crank in "democratizing the web", >> namely, opening it up from the software programmers to the >> document authors". What we have done in the XForms design >> starting in 2000 was to carefully enumerate the most >> common Web programming tasks for which authors have to >> resort to scriptig, and built the next generation design >> to obviate those explicit programming tasks. >> >> Now, does this obviate scriptin? No. Just as HTML4 in its >> time obviated the need to write Visual Basic or C++ >> programs for the simplest client/server tasks, so XForms >> eliminates scripting for most commonly asked for use cases >> on the Web; to name a few, data validation and dynamically >> growing /shrinking collections to name but a few. >> >> If and when we move beyond today's somewhat wasteful >> debate over trying to place yesterday's technology against >> today's, I believe we will have a happy web where XForms >> enables the Web developer looking to get his work done >> quickly realize his goals, whilst enabling the bleeding >> edge Web developer discover the next set of Web needs by >> experimenting via scripts. >> >> today's debate over "I can write script, who needs a >> higher level abstraction" is no different from assembly >> programmers in the early days of C compilers complaining >> about not needing anything more since they could write >> assembler well enough, thank you very much. >> >> -- >> Best Regards, --raman >> ------------------------------------------------------------ >> T. V. Raman: PhD (Cornell University) IBM Research: Human >> Language Technologies Architect: RDC --- Conversational >> And Multimodal WWW Standards Phone: 1 (408) 927 2608 >> T-Line 457-2608 Fax: 1 (408) 927 3012 Cell: 1 650 799 5724 >> Email: tvraman@us.ibm.com WWW: >> http://almaden.ibm.com/u/tvraman (google:raman+labrador) >> AIM: emacspeak GPG: >> http://www.almaden.ibm.com/cs/people/tvraman/raman-almaden.asc >> Snail: IBM Almaden Research Center, 650 Harry Road San >> Jose 95120 >> Jason> -- Best Regards, --raman ------------------------------------------------------------ T. V. Raman: PhD (Cornell University) IBM Research: Human Language Technologies Architect: RDC --- Conversational And Multimodal WWW Standards Phone: 1 (408) 927 2608 T-Line 457-2608 Fax: 1 (408) 927 3012 Cell: 1 650 799 5724 Email: tvraman@us.ibm.com WWW: http://almaden.ibm.com/u/tvraman (google:raman+labrador) AIM: emacspeak GPG: http://www.almaden.ibm.com/cs/people/tvraman/raman-almaden.asc Snail: IBM Almaden Research Center, 650 Harry Road San Jose 95120
Received on Saturday, 12 March 2005 00:49:37 UTC