Parsing - Computerphile
ฝัง
- เผยแพร่เมื่อ 1 ต.ค. 2024
- Parsing applies to human language as much as computer code. Giancarlo Sandoval on the basics of parsing.
/ computerphile
/ computer_phile
This video was filmed and edited by Sean Riley.
Computer Science at the University of Nottingham: bit.ly/nottsco...
Computerphile is a sister project to Brady Haran's Numberphile. More at www.bradyharan.com
![Building a Parser from scratch. Lecture [1/18]: Tokenizer | Parser](http://i.ytimg.com/vi/4m7ubrdbWQU/mqdefault.jpg)
"My dad is coming home"
No need to brag, dude.
"my dad is coming home" = undefined
YES!! Please make this into a full-blown series about compilers!
Michael Incog +1
i would like to see that as well
I recommend checking out Bisqwit's compiler series
Please, that would be great
NOOOOOOOOOOOOOOOOOOOOO!!!
I tried to understand this video, but I think I'll parse.
Great topic - very poor presentation (this was a „just wing it“ compared to the rest of the computerphile league...)
I was thinking exactly the same thing. This presenter was comparatively terrible
I thought that he was rushed because he had a glorious revolution to get to that he was already late for. (Also a little nervous that the government forces were about to knock on the door.)
I think he was stoned.
I think he was among the best I've ever seen on here. He described every subject with such a specificity of language as to leave no uncertainty in how we ought to understand each sentence. Perhaps you prefer the palavar of a professional personality, personally I'm happy to meander with a peripatetic if it means a precisely parsed presentation
@Matthew Explain it to me like I'm 6, then I'll believe you actually understand it. Achieve that lofty goal, and I suspect youll answer your own question.
I am EXTREMELY interested in compilers and parsing. This speaker, however, didn't seem very prepared for this video and wasn't able to ad-lib an explanation coherently. Please make more language videos, though!
Indeed. If this guy had been lecturer in my compiler course in university, I would have immediately skipped the course and started to study something like ... eh ... let's say COBOL programming.
It's so hard to build a compiler, P.S its 3 in the morning and i'm just done with an assignment on Lexical analyzer.
I don't think being unprepared for the video was the cause of him explaining it badly. I'd rather say it was a language barrier. To me, he just seems to struggle with speaking English
Yeah, it's very muddled. If you already have a grasp of lexical analysis, syntax, grammar and semantics, all words that he chucks in, then I can't see how you'd gain anything. If you don't, then he's trying to explain one thing by making passing references to a fistful of others that you probably don't understand either. Useless.
If you are extremely interested in Compilers the there is no way past THE authoritative book on the subject the Dragonbook, Compilers. Principles, Techniques and Tools from Aho, Alfred V., Sethi and Ullmann. It explains all the subsystems needed and illustrates how to implement them.
I built a recursive decent parser with that, but it is a lot of work. IMHO it is easier much faster to use YACC to get a parser implemented.
Great topic. But it needs a better explanation.
Just because someone can speak doesn't mean they know how to explain what they are thinking about saying.
And what was that?
That was parsing?
First a gloss over the Lex Analyser, then a hint at context-free grammar and then infer from that a vulnerability.
A decent context free grammer takes error cases into account as well, otherwise it cannot recover from syntax errors and the parse process has to be aborted halfa**ed.
I know what he was talking about and he still almost managed to confuse me.
I think this would warrant a few flicks to explain it properly.
I think they are trying to break it down into a simplistic manner for ease of understanding to those who have no idea how a compiler even works.
He didn't go into details as to optimization and memory handling, but it's not relevant to the basics of the concept of parsing.
Although he seemed to skim over parsing data at the higher level which I think is something that could have been discussed a bit more indepth.
Yeah... He kinda lost me when he pulled the topic of buffer overflow-type vulnerabilities from parsing ambiguity. I honestly have no idea how he managed that, that's some insane leap in logic. I mean, it's not hard to get buffer overflow in a poorly-written parser, sure, but that has very little to do with parsing ambiguity really.
@@lierdakil doesn't C++ not do that with all data types? Rarely use it these days, but I think remembering how one has to ALWAYS parse any user entered data because of buffer overrun problems.
@@RationalAndFree You seem to be confused. C++'s memory access quirks are not at all relevant in the context of a general discussion of formal grammar-based language parsing. You don't have to use C++ after all, might as well use a safer language. And certainly C++'s quirks don't themselves arise because of any parsing ambiguities, but because C++'s memory model is inherently unsafe (for historical and performance reasons).
please revisit this topic again with a better prepared speaker, don't get me wrong but even for a short video it could be better messing things up at the end... or do it with the same host just let him practice and get some advices from some of the other speakers to get more fluent on this topic and speaking about it
>Parsers are only for compilers
This is wrong. Parsing is used for many different purposes such as config files or network packets.
>Parsers must have a lexing stage
No, this just simplifies your parser. Depending on what you are parsing, a lexing stage might even be pointless to do.
>Handmade parsers are unsafe
No, using an insecure language from the 70s is what is unsafe.
As someone who has learned how parsing works not too long ago (2 weeks ago I had my exam, but failed because I made a LOT of small mistakes😅) I fully agree.
I hope this is just the first of a series, because this video was *all over the place* :(
Parsing has to start with lexical analysis, because that is the level that looks at the characters of the string. Whether there are any further stages on top of that is an entirely separate question.
@MichaelKingsfordGray The term comes from linguistics, but it is also used in computer science for computer languages. That is the sense being meant here.
Lawrence D’Oliveiro It does not need to start with lexical analysis. I can parse a number by simply looking for a string of digits 0-9. This doesn't need me tokenize it first.
@@gyroninjamodder That’s what lexical analysis does.
First time I'm actually disappointed with a Computerphile video.
But give the guy a chance. It's not always easy to sit in front of a camera explaining stuff. I'm sure it will get better with time and a little patience.
Don't just write down the things you're saying. A visual aide should complement and illustrate the verbal content, for example, you might show examples.
We made a small compiler in our compilers class. It was all fun until we had to convert to assembly language.
next time, convert to IL first, then IL to assembler.
LLVM should help with this.
Another easy thing to do is generate code for an interpreted stack machine.
2:03
no hes not :(
I was actually really invested in this but Giancarlo seemed as though he needed some practice with his delivery. Definitely bring him back because he made his description easy to follow and understand, even if his speaking voice wasn't the best.
Is the information necessary to do lexical analysis already built-in (inside the compiler)?
What happens, during the parsing process, when the syntax of our code happens to be wrong?
like forgetting to write a " : " in front of our Python "for" loop ?
Eye contact==non-existent; Computer nerd status confirmed
I'm pretty disappointed with this episode. As a computer science student, I just study compiler last semester and I realy think that without this knowledge I wouldn't be able to anderstand a word of this video. The description of the main steps of compilation was very confuse while it can be at first presented with a few general words to begin with and then eventually go into details. because here I think we never anderstand why it is like that, nor what each step realy do that the others don't finaly more important where are you going to.
Honestly, i was a bit disappointed with this one.
I was hoping they go into more detail about the basics of parsing with a direct example:
5 + 5 = 10
Parsing (into tokens):
[5] [+] [5] [=] [10]
Lexing:
[operand - 5] [operator - +] [opator - =] [operand - 10]
Syntax (based on a Grammar):
S -> [operand] [operator] [operand]
[operand] -> number | variable
[operator] -> * | / | + | - | % | !
Then it constructs an abstract syntax tree, with = being the root, and the left child being a subtree with + being THAT tree's root (leaves of 5 and 5), and then a right most leaf of the "=" tree of 10.
Then it feeds this tree, along with the lexed data into the compiler, which then generates the new output.
It would have a been a GREAT idea to touch on Deterministic Finite Automata as well, as those are really cool (only need the basic ones, not a push-down automata or such).
I LOVED making a Java 6 compiler in college. Was an absolute blast. And yes, it was from scratch. Such a rigorous semester.
I hope you would do more videos on this topic.
I don't think a layman can understand anything in this video. This video is very incoherent and not to the point. To slides off to tangents that are irrelevant to the core subject. I mean buffer overflows?? I would have expected an examples of grammars with left and right-handed recursion and how you can get into trouble with right-handed because you'd possibly need infinite stack space. That is what he should have talked about instead of just throwing buffer overflows in there. TL;DR: a poor and incoherent presentation.
It's ok, professionals are having a difficult time understanding anything this guy is saying. This is a confused explanation.
I think I've learned more through a quick google search than this video.
I am studying compiler design right now. Great timing.
This guy just threw out a bunch of words without explaining anything. Probably the worst computerphile video I've seen
I don't know this guy but I like him
Please do more of this great topic
If people are complaining about the way hes explaining....ok, then why are you on this video if you already know this topic?
Just s heads-up Sean, typo in description. I think you mean applies not apples
Fixed thanks :) Sean
Now you're just doing an applies to oranges comparison ;)
"Parsing apples to human language" - that's what cider companies do.
XD
I dont know why programming teachers always seem to make programming harder than it actually is.
Maximum irritation accomplished....
Wow, I like this explanation a lot! Thank you! And the guy looks cute also☺
Wonderful explanation! thank you~!
Poor content, don’t invite the guy back please.
This is a great explaination. Thank you
can someone explain this to me like you would a 5th grader please lol
I recognize that ancient dot matrix printer paper.
I liked it, thank you for introducing me to parsing!
Can we have Mike Pound do a video on Parsing instead, please?
Please, allow subtitles for the video... ):
BAD, BAD, BAD
Looking forward to the weird machine video!
he looks like a straight up dollar store markiplier
amazing explanation thank you
3:18 I might just be misunderstanding this, but that doesn't sound anything like the definition of a context-free language to me?
I mean, sure, if you don't know the grammar you can't exactly parse strings in it, but that's also true for everything from regular grammars to unrestricted ones?
I think he moved on and was just explaining the point of a grammar itself, not explaining what a CFG is.
that was excellent! thank you.
Grok that parser.
It's very difficult for me to focus on what the presenter is saying if they do not look into the camera (or the person they are talking to). Even watching a video, I learn better if I have the "eye contact". Perhaps that is something to work on to make the next content better :)
What did I just watch??? I was eating pickled eggs and enjoying weed...then I came to this channel and the words that came out of his mouth was english, but I did not under stand a word out of Johnny Depp's Mouth...
Why no subtitles ?
Little Bobby Tables
I feel like the only reason I understood this explanation is that I am a CS student myself having already studied this. However, I do not feel like a rooky could make any sense of this. It feels more like a recap of compilers 101 when sitting in compilers 102 than an actual explanation. For example: why mention kinds of grammars, if you are not goin to explain them?
2:04
Seven years later.......
As a Master in Computer science, having done my Bachelor's thesis on Lexical and Syntactic analysis I can safely say, that the presentation was bad. It was all over the place with details that I would never have understood without my prior knowledge. Yet the video is obviously not targeted at my level of knowledge in the first place.
finally!
Nobody has a French module because it went crazy and claims that you have 5 tokens for the number 90! (90 is pronounced in French as: 4x20+10)
Thank you so much!🙏
meh
Brian Tompsett University of Hull Languages and Compilers whaddaaap
2:10 I know this isn't a video about lingustics, so it's not super important, but "my" is actually a determiner (specifically a possessive determiner), not an adjective. Although some authors call them "possessive adjectives" (for the sole reason, I assume, of confusing people), they do not behave like adjectives, but like other determiners (e.g. "this", "the", "some", etc.).
You could also make the argument for labelling it as a a genitive personal pronoun, but given that Modern English's case system is virtually non-existent, that seems inappropriate.
BORING
I'm not sure that he really understands parsers all that well. To be fair, it is a complicated subject, especially recursive decent parsing.
I have an app ideal but don't know how to go about developing it
If I may suggest a slight correction: Parsing is not _understanding_ the string - it is understanding the STRUCTURE of the string given a set of rules describing how a SYNTACTICALLY VALID string looks. The parser's job is basically to determine if the string adheres to those structural rules or not; if it does a second job of a parser is to create an internal representation (often a tree-like structure - typically called a _parse tree_).
Anyone know of any Compiler Design books that talk about Security?
O kurczę, Gonciarz!
Don’t know what y’all are talking, but I understood him well. Great explanation.
double.Parse(ReadLine());
parsing??? i feel like you have been monitoring my watch history
If I had children, I’d name them Lexer and Token. A few years later their little sister Latencia arrives.
What camera are you guys using?
The input streams are specifically in token form, right?
I am particularly interested in parsing keywords within natural language which is difficult. I have inherited a mail Add-on that can replace variables with various things including any custom mail header. The general form is %headername% but one user recently had a problem with false positives caused by encoded URLs. A lot of the complexity of parsing can be dealt with through regular expressions but sometimes one needs additional strategies.
How about a new series just on context free grammars. You could use Pascal as an example.
I had to create a compiler for a similar to C language syntax last year at uni.
Oh boy, it was fun doing it. Super hard, but the most valuable project the uni gave me.
And i did it in Java, using java bytecode. Pain!
You look like the guy that likes to parse poems next to a fireplace on his free time.
Don't mind me, just parsing through.
Does this guy talk to real people with the same eye contact evasion that he does to a camera?
What is that around his neck ? What kind of earphones are those ?
Lol, he looks like the hacker guy from Goldeneye
I don't get the hate towards this video. Great content and presentation.
I KNEW he was italian by the first syllable. Grande Giancarlo, spiegazione fantastico e ottima dialettica ❤️
Unexpected Token "black had acters" at 5:43
Johnny Depp sure has fallen on hard times.
XHTML 2.0 when?
I like this guy's brain. More of him please
I enjoy Numberphile, even knowing little about mathematics. Computerphile, on the other hand, seems impenetrable to me
nubbinthemonkey to be fair a lot of the topics aren't presented very well. Even as a CS major who's worked with this kind of thing it's hard to follow
some eye contact would be nice
This is explained well in From Nand to Tetris, which I recommend.
Need more videos on compilers
Im doing this in Uni, it's such a pain...
6:29 Who's behind that door?
Seen one time. That's a first
now make a video on buffer
Perl is the beast use pcre lib
Seriously dude, if you're reading these comments, they don't understand how little they understand and how clearly they could. Don't ever change
Any lock parsing addicts watching this?
More.
I agree with the rest of the bunch that this video is one of the crappier ones. Not crystal clear = no fun.
I was just parsing through.
PEBKAC