RE: Declarative vs. procedural

I tend to agree. For anything more complicated than <input type="text" speech="true"> (or whatever equivalently trivial markup), the developer is going to have to start making UI logic decisions  for which javascript will be more appropriate.

From: Satish S [mailto:satish@google.com]
Sent: Friday, October 21, 2011 2:05 PM
To: Charles Hemphill
Cc: Robert Brown; public-xg-htmlspeech@w3.org
Subject: Re: Declarative vs. procedural

There are plenty of Javascript APIs that were added to HTML in the recent years (under the banner of HTML5, for a sample please see http://slides.html5rocks.com<http://slides.html5rocks.com/>) and there is no shortage of developers adopting these. These range from very simple (e.g. localStorage) to the moderately complex (e.g. canvas).

On a related note, the only use case I see for purely declarative markup is dictating into an editable field (which doesn't really need an API and can be done with a browser voice IME). To do anything more than that and to process results the developer needs to use our Javascript API. So I think it should be a non-goal for our group to produce a purely declarative model for speech recognition.

Cheers
Satish

On Fri, Oct 21, 2011 at 6:24 PM, Charles Hemphill <charles@everspeech.com<mailto:charles@everspeech.com>> wrote:
Hi Robert,

I agree.  I happened upon the JavaScript developer number and then looked
for numbers related to Web developers who know HTML and/or JavaScript.
Unfortunately, I've not found anything yet.

I did find the 2% number for JavaScript developers interesting and thought
it worthwhile to share.   What kind of adoption can a relatively complex
JavaScript speech API expect?  Is it better to have simple declarative
markup that can be augmented by JavaScript for more dynamic/complex cases?
Historically, the Web has grown successfully based on the latter approach.

It seems that the latest discussions support the later approach, but it also
seems to vary from week to week.  It's important to keep the target audience
in mind moving forward.  The developers probably don't know about speech and
they may not know JavaScript.

Best regards,
Charles

-----Original Message-----
From: Robert Brown [mailto:Robert.Brown@microsoft.com<mailto:Robert.Brown@microsoft.com>]
Sent: Thursday, October 20, 2011 2:55 PM
To: Charles Hemphill; public-xg-htmlspeech@w3.org<mailto:public-xg-htmlspeech@w3.org>
Subject: RE: Declarative vs. procedural

I'm not sure this is the right study to draw that conclusion from. The
language that's missing from the list is "HTML". It would be more conclusive
to find out what proportion of HTML web apps include java script.

-----Original Message-----
From: public-xg-htmlspeech-request@w3.org<mailto:public-xg-htmlspeech-request@w3.org>
[mailto:public-xg-htmlspeech-request@w3.org<mailto:public-xg-htmlspeech-request@w3.org>] On Behalf Of Charles Hemphill
Sent: Thursday, October 20, 2011 12:05 PM
To: public-xg-htmlspeech@w3.org<mailto:public-xg-htmlspeech@w3.org>
Subject: Declarative vs. procedural

Hi Everyone,

I mentioned some numbers from an article that I noticed last week:
http://www.tiobe.com/index.php/content/paperinfo/tpci/index.html.  The top 2
programming languages are known by about 17% of the programmers (each).
JavaScript is known by about 2% of programmers.  This might be surprising
given the billions of Web pages out there.

Having a JavaScript API is fine, but keep in mind that there are many HTML
developers who know little to no JavaScript.  It's good to have a
declarative markup option for those who specialize in markup.  The
JavaScript API should then fit with and extend the declarative markup
option.   That's easier to do now than to retrofit later.

Best regards,
Charles