Angle Brackets - Computerphile

[@stensoft]

is valid HTML, not just accepted. HTML is not XML, it's SGML and that allows tags that don't need terminating slash. But because of that you need to know what each tag means (whether it requires termination or not) to process them, and that's why XML with mandatory terminating slashes (and strict tree-like structure;) was introduced so you can easily process documents without actually knowing what's inside them, however XHTML that uses it never really caught on web pages. For anyone who deals with web pages processing, it's a pity.

[@viora_]

XHTML would have caught on if it was an official recommendation for HTML5. Really a shame it didn't.

[@ihnwtpu]

+Jan Sten Adámek In my experience with HTML sometimes it actually doesn't like when you put a slash at the end. For example, doesn't work but does.

[@SolarShado]

+Samuel Lakatoš I'm pretty sure tags aren't allowed to be 'self-closing' like that, even though (as I'm sure you know) they often don't wrap any content. It's definitely an odd quirk, though I suspect there's a reason for it.

[@rich1051414]

+Jan Sten Adámek I thought technically is the same as . When is it right to do one, and not the other?

[@TheFerdi265]

+Richard Smith AFAIK, in XML, and are equivalent. In HTML, does not even exist IIRC

[@liquidminds]

So weird. So many forgotten memories from first semesters at university... Just heard the name backus and immediately backus-naur form sprung into my mind. Thx for pulling all those memories back out of my head. I appreciate it. It's like time-travel :-)

[@paulsaulpaul]

This guy needs his own video series. He's like the Professor on the Periodic Videos channel.

[@manamimnm]

It is nice to see how we got to what we have today through the time!

[@Jacob011]

This prof. can make anything interesting. Man in his place.

[@IceMetalPunk]

Recently, a KZbinr named SethBling created a working lexer, parser, and interpreter for the BASIC programming language, all within Minecraft. That kind of programming-language-inside-a-program could be an extremely interesting topic for a future Computerphile video! (Plus, it gets us Minecraft fans hyped :P )

[@slendy9600]

i was watching one of Sethblings streams when he teased a video that would soon release which he called his "magnum opus", and that was it :D... its amazing what that guy can do

[@yash1152]

> _"That kind of programming-language-inside-a-program could be an extremely interesting topic"_ * yeah, lol. more like "hidden logical structures in programs" * includes those inside powerpoint, or ms word, Conway's game of life, as well - there have been videos on those already. * and if u move beyond programs, then logic can be implemented in all sorts things - with dominoes, with water siphons, with chemicals (dna), with electric circuits, and so on, and so forth....

[@StasPakhomov-wj1nn]

having the last line of the video be "where things finish" was a nice touch ;)

[@dark_nation_7]

I actually started writing a lexer for a language using utilizing YACC and BNF. Would love to see an episode discussing that!

[@commentsforonce7088]

04:07 for some reason I can only think of "SGML" screaming; "This is not even my final form! "

[@trailkeeper]

Maybe a good way to think of all this is as: and

[@MedEighty]

I love the Backus-Naur form.

[@shanewakeling6698]

Gotta love Professor Brailsford, hes vids are excellent

[@JL-dz8sj]

John backers was one of the big investors of 4chan?

[@Computerphile]

John Backus, FORTRAN

[@androumac1169]

Kahn Worfers, 4TrekTNG. GG myself, I think I've created a possible meme. i am dah mem loord nao. c: gg no re m80s

[@androumac1169]

I was actually trying to type Jahn for John, but iOS' autocorrect made it Kahn, so I went: "Screw it, I'll stick with this now since this actually gave me an idea.", and I basically made that reply.

[@Xclann]

Will be interesting if this series were to evolve into a talk on designing and implementing compilers.

[@hellterminator]

That is *far* beyond the scope of this channel.

[@billg4747]

I love that Prof. Brailsford is using line printer paper as a scratchpad.

[@brianmiller1077]

It's an adaptation of Numberphile's brown paper that gets written on.

[@tomlee7073]

I love the way he sings "...and you see things like, pointer bracket, 'P', pointer bracket!" at 4:41 :)

[@Cyndaquazy]

I had a class on compilers a couple semesters back. I was eating, sleeping, drinking BNF!

[@josephgaviota]

BNF, sounds like a railroad to me.

[@realraven2000]

Can you do one about the history of the => arrow operator, and how it can now be used in ECMAscript? That would be awesome.

[@wobblycogsyt]

They both use angle brackets but to claim BNF was the forerunner is going a bit far as they are fundamentally different tools. I'm sure the SGML / XML designers saw the notation in BNF , liked it, and decided to produce something that looked similar but that doesn't mean BNF is a direct ancestor of XML. BNF is much more like a DTD (or XML Schema) in that it defines what can be written not what is written.

[@lawrencedoliveiro9104]

I think so, too. The resemblance is more superficial than deep. And BNF is a more powerful notation than DTDs. DTDs are not very useful with XML anyway.

[@rich1051414]

Although the computer languages which are easiest to read are great when working on code with others, i do love the faster styles when doing something on your own and time is of the essense. Java is a language which is quite easy to read, but it does have shortcut styles available as well if you want them like: boolean flag = (i < 10 ? false : (i < 20 ? true : false)); Which is the same as: boolean flag; if(i < 10 || i >= 20){ flag = false; }else { flag = true; } The first is much quicker, but the second is easier to read. Dunno why I got off on this tangent, it just reminds me of the ever present debate of ease of coding vs ease of reading.

[@harrisonharris6988]

simple solution: make a smarter compiler to make it efficient as it compiles.

[@realraven2000]

+Richard Smith These trinaries are slightly dangerous as they have a low priority and you should use parentheses if you ever plan to nest them. You can have lots of unexpected side effects when you use them in conjunction with string concatenation. Consider this Javascript let msg = isEncrypted ? getStringBundle('warning.encrypted') : 'using Account: ' + account ? account.key : '0' + ' ' + ' identityName = ' + id ? id.identityName : 'empty'; always put parentheses around the three operands to make sure it works correctly. let msg = isEncrypted ? getStringBundle('warning.encrypted') : ( 'using Account: ' + account ? account.key : '0') + ' ' + (' identityName = ' + id ? id.identityName : 'empty'); otherwise you can get very weird unexpected results

[@HaroldSalaila]

+Richard Smith Just note that it is also possible to write that expression as: boolean flag = !(i < 10 || i >= 20); which in my opinion is far more readable than the options you provided.

[@rich1051414]

Harold Salaila yes i know that. I was giving an example of inline coding style. Perhaps in hindsight using a boolean for the example was going to lead to such corrections.

[@rich1051414]

+Realraven2000 You are right, i corrected it.

[@bernardobuffa2391]

hi. could you do a bit of history of object orientation in programming? Thanks for the videos, I don't miss one, they are all so interesting

[@Obtaineudaimonia]

I didn't think it was possible for Chomsky to go up in my estimation any further... but he has.

[@jamesgrimwood1285]

Oh god, I'm having flashbacks to the formal language part of my degree, and the terror of trying to make Lexx & Yacc behave.

[@AdamCradamParkes]

also, ::= is translatable into various human languages, i.e. ::= whereas S -> LT is gibberish in any language

[@yash1152]

Coming from "Parsing Explained - Computerphile" @ 1:29 Please optimise the description of this video for SEO - put some more search terms in it like BNF, grammer, backus naur, formal language, etc.

[@JaccovanSchaik]

Will there be an episode about lex and yacc sometime in the future? This seems like a natural extension to this one, and also to the earlier episode about state machines.

[@DuncanEllis]

+Jacco van Schaik that was exactly my thought - input to parser generators is usually based on BNF.

[@fobusas]

+Duncan Ellis (Dunx) I'm reading a book about compiler design, so this is really topical :)

[@DuncanEllis]

Vaidas Šukauskas wait until you have to implement a domain-specific language in JavaCC!

[@lawrencedoliveiro9104]

After many years’ experience, most of us have come to the conclusion that lex/flex and yacc/bison are not really worth using. Yes, you can generate automatic parsers for full LR(k) grammars, but we usually find we can get by just fine with LL(1) grammars, which can be handled quite easily with recursive-descent parsers. And hand-coding parsers and lexers makes them faster anyway.

[@brotherandy]

There is now!

[@nekture]

Oh wow. I never knew Chomsky had a history in computer science.

[@Tupster]

it was weird for me to find out later that Chomsky was more than a contributor to CS

[@TheVladBlog]

+Jason Wilkins Chomsky is life in the majority of applied artificial intelligence.

[@nekture]

I only knew him for his political activity.

[@sergioavila2720]

are you in the field of computer science?

[@nekture]

No, I am not.

[@b2gills]

The part where he said that it doesn't appear in the language, doesn't apply to Perl 6. That is almost literally valid code in Perl 6, both in the grammar sub-language (similar in appearance to BNF, but more powerful), and in the language as a whole.

[@CarterColeisInfamous]

5:20 the doctype of the html can dictate the way its parsed by the browser and if the tag may be omitted

[@lawrencedoliveiro9104]

2:23 I think an early draft of ALGOL-68 very nearly violated that tradition by introducing “::=” as a syntactic symbol. Thankfully this was changed later.

[@skit555]

So... There is a link between BNF and XML? Luv it :3

[@Nevhision2]

RIP Peter Naur

[@berndeckenfels]

Next up, roff notation for manpages (which in turn use metastatic variables in ankle brackets and cursive ,)

[@originalmianos]

He thinks "someone saw BNF and used it for SGML" is just hearsay without referring to the original authors of SGML. Interestingly people have tended to move to double braces for templating these days because they occur much less regularly in normal text, making them, although ever so slightly more verbose, more practical as you almost never need to escape them.

[@Hunnter2k3]

+Rob Fowler One character I have a love of using since it pretty much never shows up in text, never mind combined with anything else: ¬ Look at it. Wonderful character. I forgot the name. Oh, negation. Bonus character: ¦ Broken bar is nice, but can be confusing if reading it. Easy to confuse with ; or | at a glance. (pipe is worse for that since you can easily confuse it with l or I) It combined with other characters makes for simplicity when it comes to templates, encoding textual data and the like. I use things like ¬:, ¬-, etc. Or words. ¬backspace Encoding non-text characters, control characters and so on.

[@HashimAziz1]

Well, I can safely say this is the first Computerphile video I've watched and failed to understand >= 80% of what was said.

[@BodyKnight]

Professor Brailsford, you put a mark on your shirt.

[@CarterColeisInfamous]

4:50 technically thats html which is a subset of xml

[@sblop]

Please make a video explaining how this is converted to machine code, and how that is executed.

[@GegoXaren]

I love the ooc syntax... Just sayin'....: import structs/ArrayList main: func (args: ArrayList) -> Int { if (args) { for (s in args) { " " append (s) print () } } " " print () 0 }

[@ruawhitepaw]

The form of BNF I've seen has no angle brackets, but puts literal text inside quotation marks instead.

[@_simmac_]

Yes, that's extended BNF (EBNF)

[@AS-rl5vr]

I'm not sure if this is a matter of preference or not, but I always found XML/HTML very verbose and tiresome to read and write. Since both XML and LISP's S-expressions were both great for displaying tree structure, how come SXML never gained traction?

[@typograf62]

Peter Naur died recently, on january 3.

[@RustyTube]

Wait... Aren’t the identifiers in XML supposed to be in lower case, or at least case sensitive, so ... is different from ...?

[@FelkCraft]

+RustyTube No, xml identifiers are case insensitive

[@RustyTube]

FelkCraft I'm surprised. HXTML (HTML+XML) requires all identifiers (tags) to be in the lower case. I always assumed it was because of the XML portion, since the original HTML had no such restriction.

[@tamasdemjen4242]

+RustyTube XML is case sensitive. HTML isn't.

[@RustyTube]

***** I was talking about XHTML (sorry, I mistyped it HXTML), which requires all tags to be in the lower case.

[@salerio61]

I was hoping for a lot more than that :(

[@liamtennant8394]

Why is it that those brackets are definable as well as can be understood as less than greater than in mathematical terms

[@kevincozens6837]

Hmm... BNF started with elements being encased in pointy brackets? The cases of BNF grammar I've seen over the years never included them.

[@Hunnter2k3]

Then you have some languages that don't care so much about command identifiers, and just use anything you throw at it. Or a subset of the language. Whitespace, for example, just uses any visible space-separating characters, space, tab, newline character. You could embed a Whitespace program in an HTML page by using &nbsp; for webpage spacing, and now you can throw your various commands here and there in amongst the HTML markup. At least, when given a simple page. It is harder to embed Whitespace in a document using ID/class/name references since that adds a necessary space in the tags. . There are ways you can use unique tag structures to eliminate the need for IDs, but it basically means abusing the semantic meaning of HTML, which makes some people Really Mad. I think polyglots were covered before. Oh no, it wasn't. Polyglot video when? Basically 2 or more languages combined in the same document that do multiple things, be it markup, scripting, etc. You can do crazy complex polyglots with the proper use of comment tags in languages, more so if you can redefine the comment identifier. Some great examples on the Wiki page for it at the bottom. en.wikipedia.org/wiki/Polyglot_(computing)

[@Professional_Youtube_Commenter]

damn chomsky, he's everywhere.

[@ramonkokoramon8741]

good

[@marlhex6280]

wow

[@THB192]

...and SGML was in many ways far nicer, and far worse, than XML ever was. But at least it had DSSL.

[@FreeBobox592]

Hello

[@Pinkskull69420]

>Meme Brackets

[@eobrandon]

Pointy brackets? I think you mean meme arrows.

[@𠀔]

[@TheHandOfFear]

The examples he refers to are HTML not XML. They have similar roots, but they are not the same thing. I'm surprised this slipped past.

[@nagualdesign]

3:41 - Cheeky finger sniff. ^_^

[@YumekuiNeru]

they are called inequality signs REEEEEEE

[@ThatNateGuy]

This has nothing to do with CS, but can any Brits identify Prof. Brailsford's accent? I'm a Yank who's still learning and never been to the UK.

[@szymongorczynski7621]

From an NI person's point of view, it's an English accent lol

[@thefrosty1925]

I'm a Brit but I'm not sure. I can rule out a few places where his accent is from though: Liverpool, Yorkshire, Newcastle and Brimingham. But, there's that many different accents with different variations in strength over here that it's difficult for me to give an answer. To be honest, he could easily be from one of the places I ruled out; but his accent isn't any especially strong and I would probably say it's from down South, probably from Oxfordshire since it is more towards the posh end of the spectrum

[@hpesoj00]

Nottingham?

[@wobblycogsyt]

+Nate Wynd (ThatNateGuy) I would say his accent isn't clearly from any one area. It's quite soft which is generally attributed to a more well off family. Someone else guessed Oxfordshire and I would agree that's a good guess although I'd perhaps go a bit further north. It's certainly not one of the well known accents like Liverpool or London.

[@EgoShredder]

+Graham Smith Hmmm....just a wild guess and probably wrong, but.....Lincolnshire? Most likely a Southern accent as you say.

[@Ykulvaarlck]

hello

[@mensrightsedinburgh4764]

is it me you're looking for?

[@U014B]

+GamerGate Edin Actually, I was looking for Lionel Richie. Have you seen him around recently?

[@grahamfinlayson-fife73]

&lt;&gt;

[@y2ksw1]

I don't agree, but just anybody can :p

[@blackasthesky]

So they basically tried to explain to humans how to explain the computer what it needs to do like they would actually explain it to a human. Actually there was no real need to define a programming language like this.

[@Mr30friends]

.

[@U014B]

+Olegach21 💩

[@MrStrutxD]

I don't really find this interesting, could this be indicative that programming isnt my thing? or is it just something that doesnt even interest avid programmers?

[@hellterminator]

It's a bit of history. Some people find it interesting, others don't. That has nothing to do with being a programmer.

[@MrStrutxD]

***** Ok thank you for your input. I appricate it.

[@MrNateSPF]

+Anon Ops Another thing to note is that markup is not programing.

[@MrStrutxD]

MrNateSPF I understand that, but i was just talking about most of these videos.

[@MrNateSPF]

Anon Ops Why didn't you post this on a video about programing? Or more peculiarly why do you watch multiples of videos that have no interest to you?

[@NathanTAK]

Clearly, no one here has heard of Grammar-Oriented Programming.