W3C home > Mailing lists > Public > public-xg-webid@w3.org > January 2012

RE: Thank you for saving my (Peter Williams') sanity...

From: Peter Williams <home_pw@msn.com>
Date: Sun, 8 Jan 2012 07:37:42 -0800
Message-ID: <SNT143-W1859240709633441B23DB0929B0@phx.gbl>
To: <henry.story@bblfish.net>
CC: "public-xg-webid@w3.org" <public-xg-webid@w3.org>


Of couse few folks will edit HTML pages. Noone did that at the outset of the web, did they?  A few realtors will use the HTML tab in a blog post, but most will use the rich edit. A few realtors will purchase a template (with the RDFa headers), but most will not. The number of folks writing a ttl file will be zero (of course). IN realitiy, they will expect our script to "do it all" (i.e. use the blogspot API to post off the magic blog entry, and use certutil(1) to mint the cert/.p12) The RDFa THESIS was supposed to be (like the sample in the spec) that an ordinary page written up with little more than what THEY think of as some funny conventions and "CSS'ish thingies" would satisfy...to get Kingsley's linked data profile off and running (so his cloud can do the heavy lift, for a fee).   Today, I could write a script that hits the existing realty user web service (which just spits out CSV vs XML in practice, 95% of the time) and for each writes a blog post to one blogspot page. I can use certutil in script mode to make a cert (and populate n SANs) - using a vbscript MSFT already supply (for the SAN DNS equivalent). Thats 1 byte change (from 05 to 04 in said script to make it do SAN URIs....). I can easily host the MSFT cert server that certutil(1) talks to, to remotely mint a cert, and make a .p12 on windows blob store, that creates a link in the blog page that provisions their iphone, and iPad. (urr...thats COOL (to a Realtor wanting to be different to the next Realtor, competing for the next listing/commission one cube over)) The only thing I missing is figuring how then to use a URL to raise a linked data  event, so all the crawlers then crawl. Better still, Id have an "arrangement" with a particular linked data cloud that crawls for me produce fixed-stem proxy URIs, so I could put that as the second URI in the SAN collection, upfront. Or I can wait for HTML5, and the settling down period. Its your choice. tool-driven (in never ending web standardization) vs user adoption (for the public).  I could go purchase Kingsley linked data server and make my own proxy URIs (and go through no doubt a miserable software learning curve). Or I could contract his cloud, so it just works. But, until there are standards THAT "are adopted" (vs being developed), I dont. We users just wait. But I must be fair (we are the last user to adopt, typically, as actual standards are JUST SO IMPORTANT). We may be the wrong adoptee for "early phase standard". That we, here, are evidently the VERY FIRST group to EVER parse RDFa files in the wild suggests Im in the wrong place (for my particular industry, who adopt as a member of the the last quartile of adoptees becuase of the "ubiquity" requirement). in our business, there is just NO MARGIN for TIMING risk.  TSemantic web, on what Ive seen in the last month, currently fits (the killer demo was the iphone working, first time). The absolute "go flags" were: (1) it CAN be done by hand (not that one does), (2) only commodity web is needed (absolutely critical test), (3) the main concept being evangelized is ready and working (e.g. Kingsley's proxy URIs and profiles) and has been waiting in the church (for a spouse to turn up) for a while, (4) the main vendor's **delivery** of the core concept makes it *easy* to perform the heavy lift into the new paradigm, (5) the new paradigm takes something you already know (certs), and somehow turned that which is familiar but somehow never seemed right into an "aha" (once all the ducks come along in a row). I do 6 of these deep dives a year (which is why I cannot remember much about the one #7 ago. And I need to get back to my plan (doing VOIP, since CISCO is coming UP in the stack, FAST). This is the second time round for me through semantic web. The first time was with you 4 years ago (when I played with code and the intellidimension platform, building ontologies on the fly from our XML metadata, and looked at hosting our own triple/quad store in our production SQL). This is the second time, when I got to comprehend the the linked data cloud avoiding all the "deployment issues") and it has a websso mission (which is all the rage in realty), using certs (that folks are used to as mandatory thingies, for US govt purposes in the foreclosure space). But, I can wait. When its not "just right", wait (CIO mantra). One has to think like a general (reviewing the troops and their readiness ), not a corporal (who likes "to charge at 'em") or a captain (who has great ideas for the shortening the baggage train from Sparta to the Indus valley). Hmm. Those "elephants" will be use to a future general..winning all the battles while wandering around Italy (not delivering the final blow, and thus losing the war).                            Anyways, one asks: why will be year 13? (and 14?) it becuase there is always a reason to wait a bit. And next year there will be another. If I apply venture-capital logic, one goes now. I saw (at my skill level) MSFT sample code be turned in a week to produce an relying party site that could work with a webid IDP (using a standard, documented protocol). I saw it use third party plumbing doing two standards (Azure ws-fedp, openid). I saw it using commodity browsers. I saw it working with a profile hosted in the linked data cloud (on a proxy URI). And I saw normal certs minting tools put a (proxy URI) name in a cert. And then, I saw it all work on a phone, first time. With that, in VC world, you go NOW. You have hit the go button. With that you run, and you have 2 years to run. Otherwise, its dead (and is one of the 4 outta 6 that fails to breakout). The art of breakout is a mix of running fast and getting everyone to adopt. Its quite a different mantra to the web, where one wants to revolutionize (based on the ultimate paradigm shift). From: henry.story@bblfish.net
Date: Sun, 8 Jan 2012 09:11:40 +0100
CC: kidehen@openlinksw.com; j.jakobitsch@semantic-web.at; public-xg-webid@w3.org
To: home_pw@msn.com
Subject: Re: Thank you for saving my (Peter Williams') sanity... 




On 8 Jan 2012, at 08:59, Peter Williams wrote:
 
I have linked to a turrle document (if you look). Its even auto-generated (from the HTML/RDFa). It even has a describedBy link to something called linkedata (at uriburner). If its in error, let me know (Id quite like to perfect it, so I can test it really works). The point is (and Im getting bored of saying it), is that TTL cannot be published by the consumer (at blogspot). Id love that it were (but I dont know how). This is the test (nothing else). But then, to be wholly fair to Google, only at Google Blogspot can one publish RDFa. At wordpress one can publish nothing. I just have up at Microsfot Spaces (before it gave up on itself, and turned me over to wordpress - whihc failed the test). I cannot stand Google, but I have to be fair to them.  Its my nature (and my upbringing). They are the only case of consumer blogsite that works with webid as we know it today. And for that (degrudingly) I have to wward them full marks. Well done Google (Ow!) Im a totaly biase judge. But even I cannot deny the truth. I recongize the future is in HTML5 (with RDfa 1.x support), Henry. I dont hold you responsible, for the delay. Nirvana is always preferable. I just have reason to doubt you will get there. Its called "experience with crypto". Das ist in der Tasche (I dont known the french equivalent, sorry).
Well clearly WebId is not going to end up big because people publish their own RDF in turtle or RDFa by hand.  It will be big when those platforms support it directly and make the user experience seamless, so that the user can just click a button and be done with it. 
As you see there are a few of us here showing how one can make the experience simple, whilst working on the spec which we improve with feedback from our experience deploying things.

Subject: Re: Thankyou for saving my sanity...
From: henry.story@bblfish.net
Date: Sun, 8 Jan 2012 08:49:03 +0100
CC: kidehen@openlinksw.com; j.jakobitsch@semantic-web.at; public-xg-webid@w3.org
To: home_pw@msn.com


On 8 Jan 2012, at 08:30, Peter Williams wrote:
 Peter williams (known in spanish as el stupido) believed the RDFa last standard was finished ~2 years ago. we learn that , just as its finally finally-finished, and its so finished just as its replacement sort of goes live.
rdfa in xhtml was finished. 
 Im used to this, game. Ive seen it before; more than once.

indeed TLS v1.2 was also out years ago and is only being deployed now. Java 7 has it btw.
 
This is wny we in realty wait 3 more years, till the semweb is 15 year old.

There are different adoption curves. NASA has been using it and so have the life sciences for a long time. Governemtns and universities are using it to publish data. Realty may who knows be the last to use it. Perhaps they never will. 
Anyway, be that as it may. How easy would you find it to publish a turtle document and link to it? Most web sites allow one to publish images, so publishing a turtle document should not be beyond their ability.

 
When its 15 years old, Ive little doubt we will be waiting will its 18 year old.
 
Until a tooling commnity relinquises control, one waits. The temptation to hold on to a small tooling market is immense. I understand that (baving been there for 15 years, with certs). We are now are entering year 25 with certs.
 
Its that simple. The warning signs are written all over. 
 
The most worrisome (but  most human) are that W3C is now all over crypto. And, I know why. Another knighthood is on the line.
 
Only a couple of days left, to prove the case for or agin. 
 
Looking more like agin, as it comes down to the wire.
 
 
 
 
 
 
 
 
 
Subject: Re: Thankyou for saving my sanity...
From: henry.story@bblfish.net
Date: Sun, 8 Jan 2012 06:16:06 +0100
CC: home_pw@msn.com; j.jakobitsch@semantic-web.at; public-xg-webid@w3.org
To: kidehen@openlinksw.com


On 8 Jan 2012, at 05:26, Kingsley Idehen wrote:On 1/7/12 11:05 PM, Henry Story wrote:
On 7 Jan 2012, at 23:59, Kingsley Idehen wrote:

Sent from my iPhone
On Jan 7, 2012, at 5:28 PM, Henry Story <henry.story@bblfish.net> wrote:


On 7 Jan 2012, at 23:16, Peter Williams wrote:So I did it on my "programmers" web site- where Im acting unconsumer-like (built a special purpose web app).
 
And, on my blogspot account, I added the meta-link to the template. But, AS A CONSUMER i cannot set the content type (of a server).
 
So here we have a conundrum.
 
The ONLY THING THAT MAKES RDFa viable "politically" is that it works in the crappy web as-is (used by 1 or 2 billion folks). It doesnt need web upgrade, or powers that the typical consumer does not have. it is supposed to work with everyone's HTML editor (even one 10 years old) uploading to an 10 year old web endpoint.

In a few months there will be both rdfa 1.1 that is well defined in html, and a well formed semantics for micro data which is also an html5 technology. I am told that both are very similar, and for some simple substitutions micro data is rdfa too.
So here you have to be a bit patient. I think you can serve what you have as html just make sure you close tags <span></span> and that you don't use the xhtml version of <span/> .  You are living a little more on the bleeding edge here, so it's just a question of being a little bit more careful.

And we pretend Microdara doesn't exist , right?
No, I did not say that. I said we wait a bit for the dust to settle and the standards to be finished.
And I am asking: why do you think the dust has settled re. the ability to make machine comprehensible eav/spo based directed graphs from html with microdata based structured data islands? There are converters/translators/rdfizers (beyond ours) out there that do this all ready [1][2].
 
Or Peter finds a way to put the xhtml mime type up.
He won't in consumer mode. 
 We are working here for a standards body, so why should we be ignoring standards?
You tell me. 

For the moment the standard of microdata to rdf is not finished, and neither is RDFa .
But RDFa is finished enough for you re. WebID but not so re. Microdata. That's my perennial problem with this overt bias. 

RDFa 1.0 is finished http://www.w3.org/TR/rdfa-syntax/
(in xhtml).  It is the Xhtml version that we are supporting. You do too.Peter williams has a problem with HTML. He does not seem to be able to get the right mime type. He wrote:
The ONLY THING THAT MAKES RDFa viable "politically" is that it works in the crappy web as-is (used by 1 or 2 billion folks). It doesnt need web upgrade, or powers that the typical consumer does not have. it is supposed to work with everyone's HTML editor (even one 10 years old) uploading to an 10 year old web endpoint.


And my answer was just wait until the standards are finished, or use the html mime type and see how far you get. No guarantees.

Both are pretty much the same: they are unfinished. So we don't pick a winner. 

You are picking a winner by preferring one of the other. That's the problem, I repeat!

I am not picking a winner. I am trying to help Peter Williams .
Notice how in the quoted text below
The ONLY THING THAT MAKES RDFa viable "politically" is that it works in the crappy web as-is (used by 1 or 2 billion folks). It doesnt need web upgrade, or powers that the typical consumer does not have. it is supposed to work with everyone's HTML editor (even one 10 years old) uploading to an 10 year old web endpoint.

he puts the problem in terms of RDFa viability and so a long term issue. Notice that if we did not have RDFa he would be complaining how he can't do anything at all (within his self imposed constraints)

Peter was expressing a FEAR that this would never work. So I tried to answer his FEARt .
You haven't allayed his fears at all. Peter is honing into the root problem re. RDFa and Microdata. I've already told you this is a problem re. text/html squatting of the worst kind. Look at what it did to your parsing efforts, look at what its doing to Peter's consumer profile simulation. And that's just the beginning. Imagine what will happen when more time and attention challenged folks wonder by WebID. 

Yes, I agree. But then what are you going to propose as a solution to Peter Williams? Microdata? But there is not RDF mapping for it that is official, and so he would have the same problem of the project working in some servers one way, and other servers another way. 

You have no reason for not giving RDFa and Microdata equal billing re. WebID. Peter's consumer profile will work fine with Microdata as he will soon demonstrate. This is basically what we've long handled by working with AtomPub or "Cut and Paste" to blogs (e.g. WordPress, BlogSpot and others).

Please help Peter Williams by publishing his data with the xhtml mime type, and all will be done.   Or ask him to wait. 
But I agree this issue does have to be solved. Think of RDFa as being in the spec with the caveat that those specs get finished and the rdfa micro data wars get resolved. I can add a note on that subject to the spec if you wish. Peter Williams seems to be very happy to have a markup that works in html.
Another solution is to have a link from the web page to a turtle file. Perhaps Peter Williams can explore that option for us from a consumer perspective - if we understand the type of link required.
Henry

Links:

1. http://any23.org/ -- Any23 converter that makes the RDF variant of eav/spo graphs from many syntaxes that includes Microdata (even better for you its Open Source and Java) 

2. http://rdf-translator.appspot.com/ -- ditto and it used RDFLib

3. http://rdf.greggkellogg.net/distiller -- ditto and Ruby based.

 
Kingsley


Kingsley
 
Now, we have broken that rule, by requiring something other than text/html.
 
Now, as as consumer, I have no choice but to subscribe to a SEMWEB-SPECIFIC identity management system (like ODS). I cannot just use a web page.
 
Now, Im torn. For CONFORMANCE TESTING (and debugging), I dont mind such literalism. For products, I expect product to back off, insisting on only 80% compliance (trading off security, correctness for perfect compliance). I expect product to do that which works (even if the engineering quality goes down a bit).
 
 
 
 
 
 
 
 
 

 
> Date: Sat, 7 Jan 2012 22:25:47 +0100
> From: j.jakobitsch@semantic-web.at
> To: home_pw@msn.com
> CC: public-xg-webid@w3.org
> Subject: Re: Thankyou for saving my sanity...
> 
> hi again,
> 
> you now have "application/xhtml+xml" in a meta-tag [1] like so 
> <meta http-equiv="Content-Type" content="application/xhtml+xml;"/>
> 
> this is very good, but only half the way.
> you need to set a header also, currently your page presents itself to the parser still as text/html.
> 
> in a jsp i do 
> 
> <%@page contentType="application/xhtml+xml"%>
> 
> wkr j
> 
> [1] http://idweb.cloudapp.net:8080/Home/About
> 
> 
> ----- Original Message -----
> From: "Jürgen Jakobitsch" <j.jakobitsch@semantic-web.at>
> To: "Peter Williams" <home_pw@msn.com>
> Cc: public-xg-webid@w3.org
> Sent: Saturday, January 7, 2012 1:35:12 AM
> Subject: Re: Thankyou for saving my sanity...
> 
> hi peter,
> 
> one more thing :
> 
> for 12 points, you should change the content-type of your rdfa-profile to application/xhtml+xml, see here [1],
> i did the same with my testing-rdfa-profile [2] today.
> 
> at this point of time, webIDRealm uses the rdfa parser, when the content-type is text/html or application/xhtml+xml,
> but i might/will change this to only application/xhtml+xml.
> 
> wkr j
> 
> [1] http://www.w3.org/2006/07/SWD/RDFa/testsuite/
> [2] http://2sea.org/sea.jsp#j
> 
> 
> 
> ----- Original Message -----
> From: "Peter Williams" <home_pw@msn.com>
> To: "j jakobitsch" <j.jakobitsch@semantic-web.at>, public-xg-webid@w3.org
> Sent: Saturday, January 7, 2012 12:48:31 AM
> Subject: Thankyou for saving my sanity...
> 
> 
> http://tinyurl.com/7luhnnv
> 
> Your debug output gave the clue, and W3C validator did the rest.
> 
> All along it was a &, vs an "&amp;"
> 
> not only is there name/address duality, there is duality of encodings of meta-characters in the meta documents descirbing resources, bearing addresses that may be names (or not).
> 
> --
> | Jürgen Jakobitsch,
> | Software Developer
> | Semantic Web Company GmbH
> | Mariahilfer Straße 70 / Neubaugasse 1, Top 8
> | A - 1070 Wien, Austria
> | Mob +43 676 62 12 710 | Fax +43.1.402 12 35 - 22
> 
> COMPANY INFORMATION
> | http://www.semantic-web.at/
> 
> PERSONAL INFORMATION
> | web : http://www.turnguard.com
> | foaf : http://www.turnguard.com/turnguard
> | skype : jakobitsch-punkt
> 
> -- 
> | Jürgen Jakobitsch, 
> | Software Developer
> | Semantic Web Company GmbH
> | Mariahilfer Straße 70 / Neubaugasse 1, Top 8
> | A - 1070 Wien, Austria
> | Mob +43 676 62 12 710 | Fax +43.1.402 12 35 - 22
> 
> COMPANY INFORMATION
> | http://www.semantic-web.at/
> 
> PERSONAL INFORMATION
> | web : http://www.turnguard.com
> | foaf : http://www.turnguard.com/turnguard
> | skype : jakobitsch-punkt
Social Web Architect
http://bblfish.net/

Social Web Architect
http://bblfish.net/


-- 

Regards,

Kingsley Idehen	      
Founder & CEO 
OpenLink Software     
Company Web: http://www.openlinksw.com
Personal Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca handle: @kidehen
Google+ Profile: https://plus.google.com/112399767740508618350/about
LinkedIn Profile: http://www.linkedin.com/in/kidehen





Social Web Architect
http://bblfish.net/
Social Web Architect
http://bblfish.net/

Social Web Architect
http://bblfish.net/


 		 	   		  
Received on Sunday, 8 January 2012 15:38:17 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Sunday, 8 January 2012 15:38:20 GMT