please do more historic stuff like this. It is very interesting to hear this sort of thing from some one who was actually there. Also it is easier to grasp certain concepts when you know the story how they came to be. thank you very much for sharing
+Computerphile Apple pie is an application of apples, not a subset! Bramley apples are a subset of apples. Bramley apple pie is an application of bramley apples!
Wow! this series cleaned a lot of things! :) I've noted the similarities in xml html and svg syntax and allways wondered whats with xml, html, xhtml and now html5. Thank you so much! :D
I have been tinkering with SVG in Inkscape for years, when Professor Brailsford just mentioned off the cuff that the standard has Adobe PDF origins. May I suggest SVG as a future topic? :D
These markup videos are wonderful. However, to my mind, HTML is aptly named while the naming of SGML and XML is confusing. I used to work with both SGML and XML, and the thing we always had to stress when teaching was that XML and its predecessor were not actual markup languages, because languages have words (predefined tags) and grammar (DTD).
I was studying Computer Science at Cardiff in 93 and remember well the first time we saw HTML and had to write our first Webpage, little did we know where the funny little language would end up.
I was in college at that time for comp sci too but they didn't know what HTML was at that time. It was just something I ran into later on after getting a degree and you had to teach yourself by looking at existing pages online (since nothing was complicated back then) and copying their code and then playing with tags to see what happened. Never heard of SGML and I had never heard of XML until recent years.
@@invisibledave I think we were lucky they made a big thing about it being something new and interesting. Thinking back it was pretty basic stuff though.
What about a video on XSD? Think that could be pretty interesting as an extension to this video as Brailsford has now told us that there are tens of thousands of XML based tagsets. How do we work with and define these? Well, XSD is an XML Schema Definition - it is itself XML that defines a tagset and maps it to data types, essentially defining an XML-based data format. You can then send XSD's off to people and then they can put it into programs and then, when that program receives XML based on that XSD, their program knows exactly how to parse it and give meaningful data to the programmer. This is really extensively used for web services. For example, if you have Visual Studio, you can import a "service reference" which uses an XSD. So much more can be said. Please do it! :D
"If all the end tags were there, you could do a brackets matching operation, in a sense a bit like in a compiler where you match open curly brace with close curly brace. You could tick everything off and say 'That is a tree'. I may not understand what all the tags mean, but boy they match alright." Incidentially, that's strikingly similar to how I approach any advanced math formula XD
I love it! There will always be a better way to do something - and MY way will always be better than YOUR way! What I really liked was that pages that would display quite well under Netscape would not on Internet Explorer and vice-versa. Microsoft, staying true to their nature, always had non-standard ways of doing things.
And you still have separate communities fighting for how the spec should be-take a look at WhatWG (founded by individuals from the main browser vendors), which W3C tend to copy.
XHTML was a failure. It was basically HTML reformulated with the strictness of XML syntax, which nobody could be bothered to follow -- not for web pages, anyway.
Only sortof, after all, they are also maintaining XHTML5. This is the same as the basic parity between HTML4 and XHTML4. XHTML just doesn't seem likely to replace HTML anymore.
I don't understand why you have to be tolerant of missing end tags. Just because some people hand code it does not relieve them from responsibility for doing it correctly.
+UKHeliBob Because a browser that does not accept missing end tags would be unable to properly display loads of poorly written websites, and most users would blame the browser and switch to a more forgiving one. It's similar to voice to text programs. Users gravitate towards ones that understand poor pronunciation. Missing end tags were allowed at one point, so now we're stuck with it sadly.
+UKHeliBob They explained it already. If you don't tolerate errors, then people will just not use your software, and will go to someone who will. A whole lot of computing is based on the idea of error correction. And it makes a ton of sense to extend this, when possible, to anything we error-prone humans make. Especially since humans design the computers and thus can have errors even in the computer-made stuff. Trying to force humans to not make errors is a losing proposition. It would only limit those who would use your stuff. A huge reason for the growth of the web is how error tolerant it is.
+UKHeliBob It's also because of compatibility. Not everyone was a fan of XHTML, and If you were to open a website that was designed for 'normal' HTML, then end tags could be ommitted.
+UKHeliBob It depends how you look at it. HTML can be as simple as a few paragraphs of text, or as complicated as a word processor or a PDF editor. Most designers aren't seasoned engineers, not to mention simple bloggers. Imagine if you had to run a compiler just to post a message on Facebook. I do agree, however, that if you're developing apps, you should follow extremely strict rules. Then HTML is no longer just a document, but rather a whole platform, the browser is your virtual machine, and JavaScript is your assembly / machine code. It doesn't mean you have to program in HTML, you can compile your application into HTML, a special form of executable.
I think the newer MS office files are somehow XML based. I know in office 03 it was just like a .doc was a word specific file. The newer versions use .docx which is the SML version. I think they did that so that things like open office, etc can open them simpler.
+jtech0 a .docx file is actually a zip file that contains XML files. You can check it out yourself by renaming the file from a .docs to a .zip and then opening it with WinRAR, 7-Zip, or whatever you feel like.
+jesse mckeown No. DTD is standard across tag sets, making it part of the meta language, and not the application. Besides, the notation isn't "the" angle bracket notation... the syntax rules in DTD are slightly different.
+sofias. orange It kinda already was, since browsers had to handle both HTML and XHTML. I mean, XHTML with error tolerance is basically just HTML with error tolerance. Plus, who needs to verify (X)HTML anymore? Browsers now just try to display what they can, and fail when they fail. The push in browsers is to be as fast as possible, and verification of any kind takes time. Developers verify their markup using separate programs, rather than relying on the browser to do it. Having two different standards for every HTML version is just taxing on browsers, so they went for the one that can easily contain the other. HTML5 will understand XHTML just fine.
sofias. orange I don't know this for sure, but I think that it's not any faster. It just assumes properly formatted code, and as long as that is true, the speed is fine. Only badly formatted code is slower. I think what happened is that differences got smaller, to the point where they were mostly stylistic. The only thing I am aware of is that you don't have to includes the slash on self-closing tags (though you can if you want.) Technically, HTML5 says you should close your tag, even if it still works when you don't. Well, that and that HTML5 still isn't case sensitive. But, as long as you test for the correct case first, that's not really a problem.
Curiously enough, XHTML still lives on in the JSF container ecosystem, which is quite large in corporate environments even in 2020. So not _quite_ irrelevant, though one might make an argument that it should. I won't, but one might.
+John Kreitlow I don't really see him "ripping into Dreamweaver", merely mentioning it since it was one of the most popular website creation programs at the time. Anyway ripping into it would be fully justified, because the code it created was atrocious.
Reckless Roges I've had to process a lot of customer data in my job, most is CSV or tab delimited but sometimes we get XML which is less readable for determining programming requirements, not nearly as straight forward to work with and also takes a lot longer to transfer and process as the files are like 50 times larger than they would otherwise need to be. Sometimes XML just isn't the right format for the job.
Unfortunately the bits about HTML are not correct. HTML is and has never been based on XML. There was a side project that to make an XML-compliant HTML in XHTML but this has been superseded by HTML5 which is not XML-compliant. With HTML5 the parsing became standardized, tooling became better, and with the power of modern CPUs the drawbacks of an XML syntax outweigh the benefits and it will almost certainly never be XML-based. There IS XHTML5, but this is just an equivalent XML serialization of HTML5, and rarely used.
+Jason Borden I did *NOT* say that HTML was "based on XML" (see your 2nd sentence) !!!! I was at great pains to say that HTML was intended to be an *application* of SGML but it was difficult to make it so because T B-L did not write a DTD The most useful long-term outcome of the "pub in the evening" discussions was XML itself which has been an enabling metasyntax for thousands of applications. However as you say, converting HTML into XHTML to make it be an application of XML was not the panacea (!) that many people thought it would be.
Joke's on you. WHAT WG HTML is no longer defined as an application of XML *or* SGML and instead brings its own parsing rules making it a language in its own right. So while XHTML and HTML 4 weren't "languages", WHAT WG HTML very much is. Though I prefer calling it a format to avoid these kinds of pedantic discussions.
You should, just to confuse everyone, talk about ML, the functional programming language. It also ends in "ML" but has nothing to do with any of the "-MLs" we've talked about in these videos.
+Jason Hardman That would make even less sense. If you write you're defining an empty tag, so the paragraph is closed before your actual text. So your text is actually between your paragraphs and not inside them. That means any style information tied to the paragraph-tag wouldn't affect the text.
![กิโลเมตรที่ 0 เดินทางรอบโลกไม่ใช้เครื่องบิน [Around The World Ep. 1]](http://i.ytimg.com/vi/toJG9oofLFY/mqdefault.jpg)
please do more historic stuff like this. It is very interesting to hear this sort of thing from some one who was actually there. Also it is easier to grasp certain concepts when you know the story how they came to be. thank you very much for sharing
You've just got to love this professor
I could listen to Prof. Brailsford speak for hours.
+Computerphile Apple pie is an application of apples, not a subset! Bramley apples are a subset of apples. Bramley apple pie is an application of bramley apples!
Proof it's possible to still be cool at 80
+9enius It always was. And it always was possible to be dull at 17 as well!
ProfDaveB
Wow! this series cleaned a lot of things! :) I've noted the similarities in xml html and svg syntax and allways wondered whats with xml, html, xhtml and now html5. Thank you so much! :D
Nice to hear someone who was actually there at the start 👍🏽
can we get some history of regular expressions please
Does SGML have a generic closing bracket? ie to close any leaf and go to its parent, no matter the leaf's type.
Great video :-) It would be very interesting if you did similar series on TeX!
I have been tinkering with SVG in Inkscape for years, when Professor Brailsford just mentioned off the cuff that the standard has Adobe PDF origins.
May I suggest SVG as a future topic? :D
Did you know SVG is an application of XML?
These markup videos are wonderful. However, to my mind, HTML is aptly named while the naming of SGML and XML is confusing. I used to work with both SGML and XML, and the thing we always had to stress when teaching was that XML and its predecessor were not actual markup languages, because languages have words (predefined tags) and grammar (DTD).
I was studying Computer Science at Cardiff in 93 and remember well the first time we saw HTML and had to write our first Webpage, little did we know where the funny little language would end up.
I was in college at that time for comp sci too but they didn't know what HTML was at that time. It was just something I ran into later on after getting a degree and you had to teach yourself by looking at existing pages online (since nothing was complicated back then) and copying their code and then playing with tags to see what happened.
Never heard of SGML and I had never heard of XML until recent years.
@@invisibledave I think we were lucky they made a big thing about it being something new and interesting. Thinking back it was pretty basic stuff though.
Dang - I've watched all of these now and I still haven't heard him actually define any of these.
What about a video on XSD? Think that could be pretty interesting as an extension to this video as Brailsford has now told us that there are tens of thousands of XML based tagsets. How do we work with and define these?
Well, XSD is an XML Schema Definition - it is itself XML that defines a tagset and maps it to data types, essentially defining an XML-based data format.
You can then send XSD's off to people and then they can put it into programs and then, when that program receives XML based on that XSD, their program knows exactly how to parse it and give meaningful data to the programmer.
This is really extensively used for web services. For example, if you have Visual Studio, you can import a "service reference" which uses an XSD.
So much more can be said. Please do it! :D
wow yes I really hope they talk about this. I was unaware of such but it seems a key component of the whole picture
"If all the end tags were there, you could do a brackets matching operation, in a sense a bit like in a compiler where you match open curly brace with close curly brace. You could tick everything off and say 'That is a tree'. I may not understand what all the tags mean, but boy they match alright."
Incidentially, that's strikingly similar to how I approach any advanced math formula XD
I love it! There will always be a better way to do something - and MY way will always be better than YOUR way!
What I really liked was that pages that would display quite well under Netscape would not on Internet Explorer and vice-versa. Microsoft, staying true to their nature, always had non-standard ways of doing things.
And you still have separate communities fighting for how the spec should be-take a look at WhatWG (founded by individuals from the main browser vendors), which W3C tend to copy.
What would SGML people say about Json/Bson?
man, I love this professor =)
+Luis Salas Lastra Agreed. He is so easy to listen to.
thanks prof you are the best
Now what is Professor Brailsford's opinion on JSON?
I would be curious to hear Prof. Brailsford's thoughts on WhatWG and the reasons HTML5 was not created purely as XHTML5.
XHTML was a failure. It was basically HTML reformulated with the strictness of XML syntax, which nobody could be bothered to follow -- not for web pages, anyway.
What about HTML5? It isn't XHTML 5 after all, so presumably W3C gave up on the strict subset version.
Only sortof, after all, they are also maintaining XHTML5. This is the same as the basic parity between HTML4 and XHTML4. XHTML just doesn't seem likely to replace HTML anymore.
I don't understand why you have to be tolerant of missing end tags. Just because some people hand code it does not relieve them from responsibility for doing it correctly.
+UKHeliBob Because a browser that does not accept missing end tags would be unable to properly display loads of poorly written websites, and most users would blame the browser and switch to a more forgiving one. It's similar to voice to text programs. Users gravitate towards ones that understand poor pronunciation.
Missing end tags were allowed at one point, so now we're stuck with it sadly.
+fnorgen
Exactly!
+UKHeliBob They explained it already. If you don't tolerate errors, then people will just not use your software, and will go to someone who will.
A whole lot of computing is based on the idea of error correction. And it makes a ton of sense to extend this, when possible, to anything we error-prone humans make. Especially since humans design the computers and thus can have errors even in the computer-made stuff.
Trying to force humans to not make errors is a losing proposition. It would only limit those who would use your stuff. A huge reason for the growth of the web is how error tolerant it is.
+UKHeliBob It's also because of compatibility. Not everyone was a fan of XHTML, and If you were to open a website that was designed for 'normal' HTML, then end tags could be ommitted.
+UKHeliBob It depends how you look at it. HTML can be as simple as a few paragraphs of text, or as complicated as a word processor or a PDF editor. Most designers aren't seasoned engineers, not to mention simple bloggers. Imagine if you had to run a compiler just to post a message on Facebook. I do agree, however, that if you're developing apps, you should follow extremely strict rules. Then HTML is no longer just a document, but rather a whole platform, the browser is your virtual machine, and JavaScript is your assembly / machine code. It doesn't mean you have to program in HTML, you can compile your application into HTML, a special form of executable.
Interesting. I never knew SVG was related to PDF.
I think the newer MS office files are somehow XML based. I know in office 03 it was just like a .doc was a word specific file. The newer versions use .docx which is the SML version. I think they did that so that things like open office, etc can open them simpler.
+jtech0 a .docx file is actually a zip file that contains XML files. You can check it out yourself by renaming the file from a .docs to a .zip and then opening it with WinRAR, 7-Zip, or whatever you feel like.
so hang on a sec, was VRML a subset of XML then?
In fact, it would be interesting to see a video about VRML and what actually happened to it.
you forgot a link in the description!
so wait im confused is HTML a subset of SGML?
DTDs seem to be written in "angle bracket" notation; can DTD be an application of SGML? Or XML-DTD an application of XML?
+jesse mckeown No. DTD is standard across tag sets, making it part of the meta language, and not the application. Besides, the notation isn't "the" angle bracket notation... the syntax rules in DTD are slightly different.
I have a book that say HTML is an extension of SGML... Is it technically wrong?
I'm curious now that HTML5 will be considered an ML now that web components will be available.
JSON FOR LIFE!!!!!
and with html5 they decided that xhtml is basically irrelevant. :(
+sofias. orange It kinda already was, since browsers had to handle both HTML and XHTML. I mean, XHTML with error tolerance is basically just HTML with error tolerance.
Plus, who needs to verify (X)HTML anymore? Browsers now just try to display what they can, and fail when they fail. The push in browsers is to be as fast as possible, and verification of any kind takes time.
Developers verify their markup using separate programs, rather than relying on the browser to do it.
Having two different standards for every HTML version is just taxing on browsers, so they went for the one that can easily contain the other. HTML5 will understand XHTML just fine.
sofias. orange I don't know this for sure, but I think that it's not any faster. It just assumes properly formatted code, and as long as that is true, the speed is fine. Only badly formatted code is slower.
I think what happened is that differences got smaller, to the point where they were mostly stylistic. The only thing I am aware of is that you don't have to includes the slash on self-closing tags (though you can if you want.) Technically, HTML5 says you should close your tag, even if it still works when you don't.
Well, that and that HTML5 still isn't case sensitive. But, as long as you test for the correct case first, that's not really a problem.
Curiously enough, XHTML still lives on in the JSF container ecosystem, which is quite large in corporate environments even in 2020. So not _quite_ irrelevant, though one might make an argument that it should. I won't, but one might.
and now we have html5 and we have forgotten about all this.
I always close all of my tags. And then my HTML minifier throws the omittable ones in the bin again. ;)
7:25 Unfortunately, Sir Tim was seen by many to have sold out his soul over the Encrypted Media Extensions saga.
Interesting video.... ML
Of course, now we have HTML5 which is not an application of either SGML or XML, but at least has its own spec.
Wow, he really seems to like ripping into Dreamweaver.
+John Kreitlow
I don't really see him "ripping into Dreamweaver", merely mentioning it since it was one of the most popular website creation programs at the time. Anyway ripping into it would be fully justified, because the code it created was atrocious.
Yet to this very day html5 is not xml-compliant, as for example you can't have empty `script` or `span` tags.
+sillyfly html5 is not xhtml.
***** Which is mind-boggling. You'd have thought they'd learned by now.
+sillyfly They did learn. They focused on html5 because there was no point in pushing something people didn't want to use.
..oh yes.. it's a different day from the other videos..
..wink wink
Nice.
Huge :P hat off 5:08
Discuss omitted php end tags
+MattyTeare i would suggest ommitting php altogether. :P
+sofias. orange tell that to FB
I just wish people wouldn't use this format for everything these days..
Y
Reckless Roges
I've had to process a lot of customer data in my job, most is CSV or tab delimited but sometimes we get XML which is less readable for determining programming requirements, not nearly as straight forward to work with and also takes a lot longer to transfer and process as the files are like 50 times larger than they would otherwise need to be. Sometimes XML just isn't the right format for the job.
Simple Graphics,
Miserable Language.
Unfortunately the bits about HTML are not correct. HTML is and has never been based on XML. There was a side project that to make an XML-compliant HTML in XHTML but this has been superseded by HTML5 which is not XML-compliant. With HTML5 the parsing became standardized, tooling became better, and with the power of modern CPUs the drawbacks of an XML syntax outweigh the benefits and it will almost certainly never be XML-based. There IS XHTML5, but this is just an equivalent XML serialization of HTML5, and rarely used.
+Jason Borden
I did *NOT* say that HTML was "based on XML" (see your 2nd sentence) !!!!
I was at great pains to say that HTML was intended to be an *application* of SGML but it was difficult to make it so because T B-L did not write a DTD
The most useful long-term outcome of the "pub in the evening" discussions was XML itself which has been an enabling metasyntax for thousands of applications. However as you say, converting HTML into XHTML to make it be an application of XML was not the panacea (!) that many people thought it would be.
Throughout the early oughts, XHTML4 was expected to replace HTML. However that never quite took.
Huge :P at 5:08
Joke's on you. WHAT WG HTML is no longer defined as an application of XML *or* SGML and instead brings its own parsing rules making it a language in its own right. So while XHTML and HTML 4 weren't "languages", WHAT WG HTML very much is. Though I prefer calling it a format to avoid these kinds of pedantic discussions.
You should, just to confuse everyone, talk about ML, the functional programming language. It also ends in "ML" but has nothing to do with any of the "-MLs" we've talked about in these videos.
+Vincent Killion Perhaps we should talk about objective camels.
+Vincent Killion Talk about Machine Learning and MatLab in the same video
He should also talk about Marxist-Leninists, to confuse people if he is even talking about computers anymore.
You need to close paragraph tags? Huh.
+andrew zuo Just write ;P
+andrew zuo In HTML you SHOULD close paragraph tags. (Unless you are short of disk space and have space CPU cycles.) In XML you MUST close tags.
+Jason Hardman That would make even less sense. If you write you're defining an empty tag, so the paragraph is closed before your actual text. So your text is actually between your paragraphs and not inside them. That means any style information tied to the paragraph-tag wouldn't affect the text.
+Bunny83 Work from the assumption that I know exactly what a well formed HTML tag looks like and look at what I wrote again.
+Error Error
! Huge :P at off 5:08
PaRa
second
First