From W3C Wiki

TPAC 2007: Session 9: Discussion with the Director, Tim Berners-Lee

/!\ This session transcript needs to be edited. Please consider updating this entry

agenda audio


>> We have discussed a number of things today, all of which, you know, a lot of things were discussed and we believe some of the [ Speaker/Audio Not Clear ]

>> All of them do the things that they do differently.

We still do it on the ones which is the same like that.

>> We start on the cracks a the architecture and what is on the community to fix over the time.

>> That is relevant given the [ Speaker/Audio Not Clear ] today.

>> [ Speaker/Audio unclear due to strong accent ]

>> I will hand it over to Tim by saying here is the man ‑‑ [ Speaker/Audio unclear due to strong accent ] recognized not once but twice.

>> [ Speaker/Audio unclear due to strong accent ]


>> I must say when I ‑‑ given that this meeting more or less includes everything that is W3C, it was difficult for me to figure out what to talk about it.

>> I am not sure ‑‑ I am not very happy with this, because talking about cracks and mortars, it sounds like we are talking about the problems on the web.

>> But in fact most of my time will be spent on talking on the ‑‑ [ Speaker/Audio unclear due to strong accent ]

What I intended to do is take a look at the architecture.

>> When you look at the building you see the slots between the two bricks and one of those is that the building is drying out and is losing it's structural connectivity and two it has been sewn together.

>> They offer expansion which is very importantly.

>> We have had various opinions today about which are the cracks and the expansion guests.

>> First of all I really enjoyed this meeting because there is so much stuff going on and it is great being in these corridors and I want to talk to you about getting the stuff down.

Keep up the good work everybody.

>> [ Speaker/Audio unclear due to strong accent ] as I said a little bit before, one of the third of the ‑‑ one third of the W3C is among the other groups.

If you can't see the groups you can find the people and talk them in the bar or buy them breakfast or whatever it takes and find out why on earth they made a stupid decision in paragraph or section two.

>> Other groups and larger groups remember that we tend to do be the white geek male cultural.

>> There are other people out there with the applications, culturals, languages and things.

>> We appear as an organization and so we can't just sit down and right down with the [ indiscernible ], and that is involved in the working groups of the IT F for example.

>> We have to be careful for the little groups.

>> We mustn’t stop on them, because obviously that is not very nice, but in particular there is one opinion o person in the group that has the contingency that is out there.

>> [ Speaker/Audio unclear due to strong accent ]

>> If you start with the URI at the stop and you go through a number of decisions on it, you go through and figure out how it is to be displayed and what it means.

The whole reason you can use the URI on the document it is because everybody shares it on the document.

It follows it up with the URI.

>> There is in fact some information that is exchanged on that.

>> So that is the whole flowchart that you go down to.

I am going to go and mention a few of the points.

>> [ Speaker/Audio unclear due to strong accent ]

>> The first one is the content side and ‑‑ [ Speaker/Audio unclear due to strong accent ]

>> This allows for example for the server to be involved.

I can write a web page and in it I can refer to an image and later on you can replace the image and that is replaced with something that has the higher resolution options and it has accessibility options and the languages and something.

>> It is very useful, if I can refer to it in a way that is independent of the server then ‑‑ [ Speaker/Audio unclear due to strong accent ]

>> We really want the pages that we present to the videos can be view able so they can understand what we were doing and why we were thinking in such strange ways.

>> But this independence is part of the server.

If the image ‑‑ if you should look at the mind site of the referring document.

And even if the server sets up a jpeg and the document says know you should pause it as a ping or a [ indiscernible ] then ignore the server.

The architecture is, excuse me, that is it processing what it is serving and it has the right to go through and serve the images later.

That is important.

We don't change the formats slowly ‑‑ very fast, but we do change.

It allows it in the video, clearly.

It is going to be changing.

>> We are just starting down that path.

>> We don't know how the politics and the containers are going to go around the compression format.

>> That is important.

And it is a coupling joint that we can format a video for the URL and formats in the future.

>> All right.

>> [ Speaker/Audio unclear due to strong accent ] negotiation is pretty useful you can do it in the Apache box in several different ways.

You can see on the icon it don't have the ping on it.

They can both exist and the content information is one or the other, whatever the browser is happier with.

>> You should only use the content negotiation and that is either one or the other.

>> You need to by one part, by the access role of the process, you need to use that to change the access role of the file and that gave me ‑‑ [ Speaker/Audio unclear due to strong accent ] ‑‑ and it stated that it is [ Speaker/Audio unclear due to strong accent ], it wasn't actually, it didn't change it wasn't the form it wasn't the same thing.

>> We have the examples and sometimes we use that as the information.

>> But if you are looking for that HTML you are looking for that information.

>> That is not a good idea.

>> If you have the sometime information that you can send from the peer or the HTML.

Make sure it is a HTML.

>> There is a bug in a lot of common operating systems.

>> This is a new way of extending the new mind types and that is putting in the hook to help clarify the new identifications.

If you have the calendar and you download the calendars what happens?

>> So you have the calendar mind type and then it launches the calendar program and it launches it on the ICF files and it makes sure it is generated on the extensions and so that the buyer is not confused or endanger?

>> What is the problem with it?

>> The problem with it is when it fires up the calendar application, and then it doesn't describe the URI.

And what it wants to do is the describe that feed.

And that is a bug here.

>> So when you pass something along on the operating system and the A PI, please know that you put the URI on the command line, because that will help the application.

>> So when you do that ‑‑ so when that happens, the [ indiscernible ] example is that ‑‑ that is when they tend to use the web ‑‑ [ Speaker/Audio unclear due to strong accent ]

>> That is only to get around the box.

>> And the complete URI scheme.

And I haven't talked an 't the URI schemes you shouldn't use them really and invent them.

>> We have heard a huge amount about it today.

>> When you are fine tuning it we don't want to offer you want.

>> Renaming it doesn't really help.

>> One of the big questions is the tactics.

I am not going to go over that in great detail.

My feeling about the taxes is that it is not ‑‑ tactics it's okay to end up with two codes.

>> [ Speaker/Audio unclear due to strong accent ]

>> I guess I find this frustrating that we haven't ‑‑ frustrating that we haven't managed to make the path on the code.

>> That is easier and I like the extensibility ‑‑ [ Speaker/Audio unclear due to strong accent ]

>> Now does that ‑‑ [ Speaker/Audio unclear due to strong accent ]

>> We will never know these things or track those.

>> I am happy with the actual modes that you are documenting on the practice.

I think it is important to find a way of moving from the messy situation to a cleaner situation.

How can we do that?

>> I think have a lot of it has to do with the Synchronization of how it works.

They can do more of it if the valid dater is within the java.

>> The idea of the XML was the X swapping was the extensibility.

You can put the new tags in when you invent the table and then we put those in and this will display the date and it wouldn't look as pretty.

>> That sort of thing would happen with the HTML.

>> That is whatever we have the value ater.

>> And that is part of the attributes.

And that is healthy.

If you put them in the HTML, then obviously you need to put them on the working group or maybe you don't need to do that at all and maybe you can have a machine, a machine on the virtual hallway and reflect which tag you want and then you got it.

>> But seriously, you shouldn't swap other peoples name spaces.

But when you use them and make your own name spaces and then you make a tag you are not treading on someone else’s shoes.

>> So how about a validater that defines in there the new attributes with the name space.

Find something there where it mentions that it is valid.

>> I think what we should do as the community is build the validations and that will take anything.

>> Compare it with the ‑‑ [ Speaker/Audio unclear due to strong accent ]

>> Why don't we validate the web architecture and we need to check on the mind sight.

We have the reason of the payments and go out and check the page and the reasonable mind sight and so on.

We have to check all of the things with the traps that are built in.

And that is something that is not and hasn't been designed in an extensible fashion.

>> So there is one way by which we can move people.

>> So you find the website and they are moving the tools.

Maybe also we can move actually how to stop giving up on it.

>> Another one, which I am talking about when I am giving talks, a classic example is if you look at the list of the top universities in the United States you go to their home pages you will find that they have home pages.

>> So we could ‑‑ [ Speaker/Audio unclear due to strong accent ]

>> We could as a community do a little bit more of the ‑‑ [ Speaker/Audio unclear due to strong accent ] when we go to give talks at the companies and run their website and validate who is the web master.

One of them should validate the home page.

>> Clearly you can run it on the whole site.

>> Check it out you know.

>> The save as.

That is the save as phenomenon.

>> Look mom I have a web page.

So when somebody does a web page you can slowly get the bugs out and filter the web.

You make it different to get to the actual source without it being solely annotated.

So if you want a save as, it a slave clean, and ‑‑ save clean and that is the [ Speaker/Audio Not Clear ]

That is the expansion and general ‑‑ [ Speaker/Audio unclear due to strong accent ] they can include the HTML tidy functionality, and maybe that would ‑‑ maybe that would help us fix it.

>> Okay.

>> We talked about the next thing quite a lot on the panel, so I don't think we should go through that.

>> As I mentioned we have discussed that on the panel so I won't go into them.

>> We have a entry return instead of using the URI's it works for the global set of terms but not the industry wide set of terms.

This doesn't allow the set of industries or a smaller community to add the terms, and maybe ‑‑ [ Speaker/Audio unclear due to strong accent ]

>> We have to be aware of the ‑‑ for give me for interrupting, but the captioner cannot keep up with you.

>> And I have been talking so slowly.


>> No.


You have not.

>> What are you going to do fire me?

>> Let's talk about the new tags and ‑‑ [ LAUGHTER ]

>> When the validater complains this causes people who design the micro formats not to use attributes.

>> What do they use?

>> They overload classes.

>> Those classes have been used for CSS.

>> They work fine for CSS.

>> Because the person who writes the HTML page also writes the CSS.

>> That person coordinates that the same class name is in the CSS and the HTML.

>> When micro format designs or uses classes it is not a local thing.

>> They have to go to Tantek or to Tatek's wiki and write down in a special wholly place and write the name of the new cause.

>> I take it back for possibly having time for questions! [ LAUGHTER ]


>> So as I said, XML was supposed to be extensible.

>> In the future we should feel that when we find a tag with a name space, one that we don't recognize, but where there is a name space document, that is not a bug.

If we find a tag with no name space, a tag that we don't recognize, maybe it's a bug, maybe it's a misspelling.

Maybe it's a misspelling that the web designer really wants to be shown.

>> Maybe it's an extension.

>> No way to know.

>> So let's talk about [ indiscernible ] versus SVG.

>> [ indiscernible ] people put your hands please.

>> SVG people put your hands please.

>> Both, put your hands up please.

>> Okay.

>> I guess I see something that is as well developed as SVG with a lot of test cases and a long history and a lot of implementations sitting there with the browsers which accept it as a format which can be just pushed in and I see an alternative in it I have to say that I sigh.

>> Next I am going to talk about the processing of the XML.

>> XML is tree structured.

>> XML is as a tree structured language normally you would expect to be processed from the top down.

All three structured things all of the ‑‑ [ Speaker/Audio unclear due to strong accent ] ‑‑ from the top down.

>> As we go down the tree, we get ‑‑ every time we go into one element and we look at the children we get another flexibility point.

There is a flexibility point when we first look at the document and if we only know it is XML, we have not been told that it is text, HTML, or applicant written, or HTML or ‑‑ [ Speaker/Audio unclear due to strong accent ]

Then we look at the first, the top element, the document ‑‑ the root element of the document and we look at the name space and it is HTML, and then we figure it as a HTML document.

And if it is a RDF, we figure it to be a RDF document.

>> And if it is SAG then we figure it is SAG.

>> And then within that we are surely working towards an architecture where we can completely mix these things and we can find a SAG with HTML and have within these things, and not only that we have the XML encryption name space.

We may find something that is encrypted and I like the architecture where we are effectively doing a lazy evaluation top down.

So when the browser needs to figure out what is inside the child it looks at it and if necessary decrypts it.

>> And that is to replace that element with itself decrypted.

>> By the way, I won't go into details now, but XFLT, also when it is in the template mode it works in the similar fashion.

>> You have the document which is almost HTML again, and inside there is an element which is a little bit of a piece of it.

>> And I favor in the way of looking at the XML document that is defined from the top down.

>> Now, in fact whether we process the documents we often just have a pipeline.

>> We have something that will do all of the XML processes and we have something else that will do the decryption and we have the long arguments that of which it will do on the XMLT and the ‑‑ [ Speaker/Audio unclear due to strong accent ] ‑‑ and X include is another one.

>> Should I do the ‑‑ should I do the XLT before the include.

What happens if I am including XLT.

So I feel the top down approach leads to sanity here.

>> But it may lead to this one, the architecture.

>> Everybody knows what the XML document means because we are all starting at the top guys.

>> When I talked about the XML top down and dispatching.

We are doing the dispatching between the two different pieces of software and that is possibly dynamically and installable at the element level and not at the document level.

>> Some people think this is totally obvious, but it is really, really stupid.

>> That is life.

>> Multiple name spaces on the ‑‑ [ inaudible ] ‑‑ is just mentioned briefly, all of the ID F and all of the link data out there.

All of the ID F saved name spaces and it works.

It is being matched together in places.

>> It is without the links in the HTML communicate.

>> So for example, if you go to data about ‑‑ this is data on the dynamic web about me, for example, a lot of this data, and it comes from the Wikipedia and it is stuffed with different languages and it is stuffed with the stuff that is being pulled from the [ indiscernible ] database and the computer science Bibliography.

And I have written that myself.

>> There is all kinds of data as you follow through this class.

>> You will find your way off, from going from one application to another application, you are putting in the data from all different sites and Charles and Kathy [ indiscernible ] have 113 friends so now we are on the inoperable sites.

>> There is all kinds of different data for extending those applications.

So it basically works and it is the plan and it is very sound.

>> I am sorry, some people don't like it, but it is absolutely core to the way that audio works.

>> So we are running out of time.

I just want to mention a few things that I think are exciting that are happening.

>> There are more things than what is on this list by a large number.

>> Semantic data and that is data when we use the ACCPT that we use in URI's so you can look them up.

>> PDF, I think that ‑‑ SVG has been around, but now that silver light of the hair in fighting for the proprietary space, so maybe we should wave the flag that we think we have a strong basis for that contender on the open standards.

>> One of the things that I am very excited about recently is down countering ‑‑ countering spam and that is within the ‑‑ spam and that is within the area and that is a friend of a friend on the net working.

>> So example on my blog if you are a friend of a friend of anybody in the group you can comment on the blog.

So that is a social net work to establish who people are.

>> Obviously we have a large number of people here and all of you have registered and but know who you are and ‑‑ we know who you are and we can use you so we can use anybody that is a friend of a friend and they can change the wiki comments on the blog.

>> [ Speaker/Audio unclear due to strong accent ]

>> That is a practical and exciting way to make the web a better place.

>> Obviously going mobile, [ Speaker/Audio unclear due to strong accent ], so there are some exciting things.

Some of those fit together, and here is a picture that I will show you.

>> Here is an example of ‑‑ underneath there are all kinds of ways on the bottom layer.

We have the existing data and that is in the HTML data and this is data integration and that is to provide all of the existing data in the HTML to virtual audio pages.

>> This blue bus is the automatic web browsers.

>> So that bus is the semantic web bus and people on that bus can expose data [inaudible].

>> There was a blog that went up a few hours ago, that said just give the data now and give it to me ‑‑ just don't do it wrong, just give me the data.

>> Give me the data raw.

>> This is if you would like is the raw data.

And this ‑‑ this is the raw data.

>> This is the economy.

These are the people like gov track which allows you to search for ‑‑ search for people ‑‑ and track their voting ‑‑ track their voting records in the senate or Congress, it's using datatran and it got a hold of the data some how and they have made a nice user interface in to it and they have made it available in RDF for everybody to see and this is the opening of the data link products that you can find for example from the ‑‑ [ Speaker/Audio unclear due to strong accent ] ‑‑ and you can look for the one particular one and find out where it was from between these dates and you can go through and see these things and do all kinds of different things within the match up.

I am doing a similar one in UK and they work for you.

>> Many of them only use data from one source and they scrape that source.

Once they connect in the database you can access all sorts of things and you do one site all across the country.

And you can do one site all across the town.

>> When you do that you can use one bus, that is HTML, and you present that to the user.

>> I think that is an interesting model for the future.

This maximizes the reuse.

It maximizes the data or obligation and lets the people see the data to put that up there and do the manipulation of it.

>> The architectures that I will see more and more in the future.

>> Thanks for listening and keep up the good work.


[Audio Cutting In and Out]

>> We may have time for a few questions.

>> [ Speaker/Audio unclear due to strong accent ]

>> I know it is tiring listening to someone that ‑‑ speaker in audience is faint and unclear.

>> Eight years ago or seven and a half years ago, I believe we made as part of the community we made a correct decision in saying that XML on the web would not actually be on the ‑‑ [ Speaker/Audio unclear due to strong accent ] mind sight but on a different hindsight, on the 20/20 hindsight, I think the ‑‑ [ Speaker/Audio unclear due to strong accent ] ‑‑ was wrong.

That is the position that could be causing the significant problem on the racks.

>> They are using that as the opportunity to go back and help fix that rack.

>> That is really complicated.

>> The question is about ‑‑ that is ‑‑ I was going to save that question for somebody because it is an awkward one.

>> You are saying that the XHTML and the HTML should be saved.

My thought at the moment is that based ton philosophy that we should have a smooth ‑‑ that we should encourage people when they serve as the text on the HTML to serve it as the fact through XHTML.

>> They should fix it up tidy through the web browser.

>> For some reason, years and years and I understand ‑‑ have not under stood that their application, the XHTML.

>> [ Speaker/Audio unclear due to strong accent ]

>> So meanwhile we should encourage the browsers to recognize all of the mind types in the W3C specks.

I think we should have a best practice implementation in the mind dot type file types and we should change on any I SP's which serve out public accounts where they don't allow you to serve W3C recommended formats.

I know lots of groups have this problem and lots of them have said [ Speaker/Audio unclear due to strong accent ] ‑‑ that is the only thing that the stupid Apache can understand.

>> I think it is good in this case.

People should do it, and maybe you could make small go around dramatically ‑‑ [ Speaker/Audio unclear due to strong accent ] ‑‑ moving the people towards the efficient marketing is the thing to do and making ‑‑ and as I have said have the validation of going completely clean to XHTML but to do it gently.

[Audio Cutting In and Out]

>> Give him a round of applause.


>> Thank you.

>> Will you be at the reception for a little while?

>> Yeah.

>> Okay.

>> Great.

>> All right.

>> Thanks. We are at the end of the program. You must be exhausted. I have a few announcements and thanks to give and the incentive for you to bare with me is I will open the bar a little early. So just bare with me. We have a nice plan up on the 16th floor you should have a nice view and a great chance to talk about all of the wonderful things that we have discussed today. It will start at 6:45 instead of 7:00. Please fill out the link on the home page. Please fill it out. This is the first time we have had this and you can fill it out now and come back to it and complete it at the end of the week if you want to.

>> I will send it out to the register ants if I can.

>> This includes the chairs on the architecture groups. This will be at 9:15. The rest of you will have the inform meetings.

>> I want the reps to attend that session and we will have a little bit of time to reflect on the consortium of today.

>> The audio and the transcript will be available and they are going to be cleaning it overtime. They have been following and there has been an ability to follow what the transcribers are doing out in Caption Colorado.

>> That is a tough job. And I would like to give them a round of applause.

>> Thank you.


>> I would like to also think the round of panelists today. Give them a round of applause.


>> We had a program committee and we also have a program committee for things like this, and can you raise your hand. I will name off the names quickly:

>> They did a great job and I hope you found the program interesting and stimulating and let's give them a round of applause.


>> And last but not least, I want to think the people who have put together, not only today, but as a week as a whole. But we have inside who is [ indiscernible ] and

[ indiscernible ] and they were joining Suzanne [ indiscernible ] and Amy

[ indiscernible ] and that is not only today but the complicated program that we have on the logistics attend the right room and the AV and just give them a big round of applause.



>> There they are there the back.

>> Raise your hand you did a great job.

>> Thanks again to our sponsors.

>> On behalf of [ indiscernible ] Cisco, and Nokia, IBM opera and so give them a round of applause.


>> We will end the session and see you at 645 and the meeting is adjourned.

>> W3C conference call adjourned.