- From: Kazuhiro Kitagawa <kaz@w3.org>
- Date: Thu, 16 Aug 2001 23:07:40 +0900
- To: www-mobile@w3.org
If you are using servlets you don't need to parse the request. The servlet is passed a request object that contains a parsed version of the request. However for CC/PP it will be necessary to do further manipulation on the request object in order to obtain the device profile. There is no standard API to do this yet - especially as it depends on the CC/PP protocol being used and that is not finalised. For an example of processing a request header in a servlet, see Snoop Servlet at http://triton.towson.edu/~schmitt/server/servlet/ Mark Butler HP Labs Bristol -----Original Message----- From: Sandeep K Gain [mailto:sandeep.gain@riverrun.com] Sent: 16 August 2001 11:38 To: 'Butler, Mark' Cc: www-mobile@w3.org Subject: RE: is CC/PP worth using its a very helpful info. Is there any open source free parser(in Java) available for the same. thanx Mark. sandeep. -----Original Message----- From: Butler, Mark [mailto:Mark_Butler@hplb.hpl.hp.com] Sent: Thursday, August 16, 2001 4:07 PM To: 'sandeep.gain@riverrun.com' Cc: www-mobile@w3.org Subject: RE: is CC/PP worth using Sandeep M-GET and M-POST are mandatory, whereas GET and POST are optional. Therefore I'd expect most browsers to use GET and POST as users will prefer browsers that try to retrieve content even if it cannot be adapted for the device rather than not display content at all. Current servers do not respond to M-GET. For example Tomcat replies HTTP/1.0 501 Not Implemented Content-Type: text/html Servlet-Engine: Tomcat Web Server/3.2.1 (JSP 1.1; Servlet 2.2; Java 1.3.0; Windo ws 2000 5.0 x86; java.vendor=Sun Microsystems Inc.) <head><title>Error: 501</title></head> <h1>Error: 501</h1> <h2>Location: /ccpp/headers/index.cgi</h2><b>Method M-GET is not defined in RFC 2068 and is not supported by the Servlet API </b><br></body> So yes this means that servers need to parse the request in order to determine if CC/PP is being used. Mark -----Original Message----- From: Sandeep K Gain [mailto:sandeep.gain@riverrun.com] Sent: 16 August 2001 09:55 To: 'Johan Hjelm'; sandeep.gain@riverrun.com Cc: www-mobile@w3.org Subject: RE: is CC/PP worth using thanx johan for your valuable comments Just one more question , if you don't mind :) Is it expected from the origin servers that they will parse for profile and profile-diff HTTP extension headers , even when a GET or a POST request is sent , instead of the M-GET and M-POST as in the HTTP extension framework? In brief , how do the origin servers know that CC/PP is being sent(in absence of HTTPex) .. May be its a very trivial question , but still ... thanx & regards, Sandeep -----Original Message----- From: Johan Hjelm [mailto:johan.hjelm@era-t.ericsson.se] Sent: Thursday, August 16, 2001 1:17 PM To: sandeep.gain@riverrun.com Cc: www-mobile@w3.org Subject: Re: is CC/PP worth using Actually, I was just made aware that the WAP UAPROF spec, chapter 9, contains just such a protocol. You will have to look it up, it looks really weird when I try to copy and paste from Word. But it looks like it takes care of Marks comments. As for the future of CC/PP, I am of course biased. Yes, we will rule the world and eat the cake, too. However, all legacy servers will never implement all new technologies. That will never happen. So you can not expect total penetration. As an aside, I am increasingly convinced about the value of the approach using a default with overrides. Of course, you need a structure like CC/PP to keep stuff in. But I participate in discussions in lots of other places, and the approach keeps getting validated. Concerning the vocabulary, the answer is "maybe". The attributes are there, and things like screensize and colour are not specific to WAP. However, you would have to look at it yourself to say whether it covered all your needs. I was just in a teleconference with another group in our company, who is adding a set of attributes in a namespace of their own in a standard they are producing (you may find out when it is finished). So that is a viable option, but I agree with Lalitha that you need the support of an organization for it to be accepted. Hope that answers the question Johan Sandeep K Gain wrote: > folks, > First a general background on what I am up to. > > I am developing a server that generates content according to the HTTP > User-Agent header information. > We have got our own browser sort of application running on Palm and > WindowsCE based devices . > The application running on these devices request content from my server and > I generate content according to them . The content sent to them currently is > XHTML. > In future we wish our application gets content from all sites on the > internet. > Now we wish to get hardware/software/application specific information from > the client and then generate the content according to that . > so far , its fine . > > Now, if we implement CC/PP , we will be able to have our application > runnning , as the server will understand the CC/PP information. > My question is should we really use CC/PP , becoz if all that is to be done > is just to pass this information, then we can pass this in our own protocol > (which we have). > Is CC/PP here to stay , i mean are we anticipating that most of the internet > sites will gradually become CC/PP aware. > Can CC/PP be implemented over HTTP without using the HTTPex framework? > And is UAProf vocabulary sufficient , or is it just for WAP applications.I > mean, I wish to use UAProf vocab for CC/PP from my Palm and WindowsCE based > applications , can i do so? > > your insightful comments please > regards, > sandeep -- ++++++++++++++++++++++++++++++++++++ Johan Hjelm, Senior Specialist Ericsson Research Japan Read more about my recent book http://www.wireless-information.net ************************************
Received on Thursday, 16 August 2001 10:08:22 UTC