Where GREP Came From - Computerphile
HTML-код
- Опубликовано: 21 дек 2024
- Commonly used grep was written overnight, but why and how did it get its name? Professor Brian Kernighan explains.
EXTRA BITS: • EXTRA BITS GREP from E...
Inside an ALT Coin Mining Operation: COMING SOON
Unix Pipeline: • Unix Pipeline (Brian K...
/ 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
At an interview -
recruiter: what would you consider your greatest weakness?
me: I'm not Ken Thompson
I'm not Clay Thompson. Oops, wrong interview
Hired.
@@u.v.s.5583 I'm not Clay Thompson, but 20 dollars is 20 dollars.
My greatest weakness is i do not think i have any weakness.
Absolutely the best 🤣
Students: this project is impossible
Professor Kernighan: KEN THOMPSON built this GREP in A CAVE with a BOX OF SCRAPS
Students: but.. we're not.. Ken Thompson..
@@Fromatic Well, no one's perfect, except for Ken Thompson
@@Fromatic Exactly.. Bruce Lee used to do pushups with on single finger..
Iron man reference isn't it?
@@Fromatic Professor Kernighan: And that's why you have a week to complete the assignment and not just a few hours.
Not being Ken Thompson is a struggle every working software engineer has to contend with.
Except Ken Thompson.
nope, he's retired.
cbernier3 - Last I heard he was still doing some work for Google.
They say he retired from Google.
hmm jeff dean disagrees
It's great to watch and hear legendary computer scientists talk about such fundamental and ubiquitous tools.
Hello my cousin.
grep is not only a program, it has become a verb. It is common among computer people to talk about grepping for something. Which may or may not actually be done with grep.
My personal favorites are vgrep and vdiff, just to specify you're doing it by eye.
diff?
"make your code greppable"
i recently asked some friends if there was a way to "grep through messenger history"
What i like most at ebooks is the fact, that they are greppable.
This video is such a gem! Computerphile is making a great contribution to the history of computer science.
1
Indeed.
hwabag
“And of course they all had one disadvantage, None of them were Ken Thompson”
Oh My!!!! Best line of the series.
both bill gates and mark zuk got that assignment.
(fake history)
@@ylstorage7085 Bill Gates is a genius, just not in programming.
"grave" disadvantage!
I've been watching videos from this channel for a pretty long time and I just now realized that this guy is THE Brian Kernighan.
/applause
But your avatar is very fitting.
Yep. For those who are still confused about who THE Brian Kernighan is, checkout K&R's C Programming Language: a.co/6C40vKc.
a wild kripp has been spotted!
Same. Mind blown.
Oh yeah, same. Someone mentioned Prof. Kernighan in the comments, I did a double take, then slowly realised who's talking in the video.
In Swedish the word "grep" means "pitchfork" which I've always though fit beautifully. You have a stream of stuff coming in and you jam your pitchfork in to grab what you're interested in.
I always mentally converted it into the dialectal "greppa [tag i]" -- roughly synonymous with "seize hold of."
In Slovak it just means grapefruit lol
In bad Dutch English, grab is pronounced as grep. So mentally I've always thought of grabbing text from input.
Like grepping a needle in a haystack!
In Norwegian we also have "Å gripe", meaning "To grab", which sounds very close to grep
Thank you Mr. Kernighan and all your contemplates like Bill Joy, Dennis Richie, Ken Thompson, and others for giving us the greatest computing environment imaginable. Amazing how after 50 years it's still so relevant because of your genius.
For context:
#programmers to develop IBM OS360 = ?
#programmers to develop Windows / OS/2 = ?
#programmers to develop Lisa = ?
and then...
#programmers to develop Linux = :) ?
"I was teaching at Princeton as a visitor, and I needed an assignment for my programming class. And I thought "Hmm!". So what I did was to tell them - the students in the class: "OK, here is the source code for 'ed'. It was at the time probably 1800 lines of C. "Your job is to take these 1800 lines of C and convert them into 'grep' as a C program. OK, and you've got a week to do it". And I told them at that point, that they had a couple of advantages. First, they what the target was. Somebody had already done 'grep' so they knew what it was supposed to look like. And all they had to do was replicate that behaviour. And the other thing is that it was now written in C. The original 'grep' was written in PDP 11 assembly language. And of course, they also had one grave disadvantage: None of them were Ken Thompson."
8:58 - 9:45
Hahaha savage
@Joshua Murphy converting to grep?
@Joshua Murphy g/re/p was already a command in ed, which prints(p) every occurrence(g) of the described sequence(re) in the document. So basically what the students need to do was to load everything on the file system into ed? Or make ed run outside of its boundary of one single document?
A very great project tbh. So clever, so tricky, disastrously hard without the knowledge of the history of grep. But without a doubt a very great project.
@Joshua Murphy Ken just read the file in chunks, did a match on a chunk, stored the results, and then did the next chunk of the file. I guessing this was a one-hour assignment for someone who knew C.
@@ruadrift not with memory limitations at the time
@@陳秉軒-c9b so the trick was figuring out that grep was just g/re/p ?
A lot of the aspects of ed that are covered here still work in vi (or even vim).
Benjamin Geiger
Vi's source code traces back to Ed. After Ken Thompson wrote Ed, Em was created (Ed for Mortals), then Bill Joy wrote Ex. Yup just like Vi's ex mode. Ex became Vi, Vi became Vim and the rest is history
Faris Chugthai Don't forget cousins like CP/M's ED, or IBM's XEDIT.
They can also work directly on the command line in BASH x]
Then Emacs came along, heralding the great editor wars that spanned three decades and saw many software engineers go off to fight in the trenches of Usenet and never return.
@@tomihawk01
And in the end, nano won.
(me starting a new war. SCNR)
This guy is one of the living legends of computer science! I still have his book at home
I keep it as a priced possession
The awe with which you, Mr. Brian Kernighan, speak of Mr. Ken Thompson... It just drives home the fact that we all just stand on the shoulders of giants. You might have received astronomical amounts of these, but here is one more: THANK YOU.
From my beginning of learning Linux and Unix, I've known that grep stood for global regex print, but never heard the full story, and it's relation to ed. Really great video!
Those ed commands can also be found in vi and sed.
AFAI grep stands for Global Regular Expression Parser. but who knows? 🤷♂
@@jcf20010 i love standardization
Always awesome to hear Brian Kernighan speak, thanks for doing the interview!
I had the honour of meeting BK when he was invited as a guest speaker to GCHQ in the early 90s when I was an intern. What an absolute legend.
It's amazing he can make it so accessible.
I learned Fortran in the mid 60s. It required punch cards and a single error would reject your whole batch.
I knew then computers would never be very important.
xkguy, … and you were right.
Yes(from FORTRAN IV in the '70s), and the complier would didn't say error but rather errata F(fatal), I(informative for non usasi) or N(for non fatal error). Submit your cards and hope for the best. I almost forgot that sometimes when formatting the printout hollerith code: "at least one redundant delimiter has ocurred".
I also learned Fortran in the 60s. I ran my first program on an IBM 360. After it worked alright, I took a dump of the memory and examined it trying to figure out how the source code was translated into the machine code. The first Fortran command was read a card. It was very difficult to find the equivalent machine code instruction in the dump. There was so much overhead.
You learned FORTRAN from John Backus? 😮
Every day I don’t use FORTRAN is one I treasure and reflect on my gladness for having the privilege to have done so
And this is coming from a guy who mainly writes R
You don't see this quality content on TV.
There's no naked women, beer, or men acquiring brain damage. In America, that means audience fell asleep.
I don't even own a TV.
...unless you cast RUclips to your TV.
Reminder that Computer Chronicles was a TV show in the 80's.
SmartWarthog so... 30 years ago? Not now then lol
...and the original grep was written in PDP-11 assembly language...!> Overnight.....
@jqbtube The fact that it's in assembly language makes it as impossible.
@@johnyang799 Not if you know assembly language.
@@jtgdyt2 Agreed. I found it to be a huge advantage in latter years, that the first language I learned and used for 2 years, was assembly.
I usually complain about the string handling in C. I don't think I've ever even THOUGHT of the possibility of doing text processing in assembly ;D. What a headache ;D.
grep - simple magic. In 2007 at the age of 60 I was applying for a job in IT at SEPTA - the transit agency for the Philadelphia area. At the time they were using networked SCO Unix systems for the commuter rail dispatching system. The first question asked during my employment interview was: What is 'grep' and how can you use it?
Having had been introduced to UNIX in the mid-1980s - it was a rather easy question.
having had been introduced?
Did you get the job?
So you nailed the interview and now you're the owner right?
@@markgreen66 Yes I did. I was most surprised because of my age. I stayed there until turning 68.
@@santiagoserna4 Owning was not quite possible as SEPTA is a rather large multimode transit agency. But I did "own" my position for a solid 8 years.
The OG (original greybeard) is back! I always, always love listening to Brian Kernighan. I admit to not knowing that ed should be pronounced eee-dee. One cool thing about ed is that if you master it, then sed and awk are a piece of cake. Like all things Unix. Thanks so very much Computerphile for having him on again. More, more please!!
When he said "25 years ago" my first thought was "1975". Then he said "1993", and I realised that we're already close to 20 years into the 21st century. Wow.
@@new-lviv This aged quite well. 2020 has been quite roaring
You guys are so 2010s
@@kebien6020 2022 has been even more roaring so far.
Very soon 1975 will be 50 years ago! That's crazy.
I'm a simple man, I see a video of Brian Kernighan and I click it.
The amount of time grep has saved through the time it has existed is staggering!
Computerphile is undoubtedly one of the best channels on RUclips!
This just blew my mind. I just started out trying to wrap my head around vim as an editor and this video has already helped contextualize so many of its commands and shortcuts. Apparently if you come out of the UNIX-World (I dont!) many many of these supposedly hard to memorize command-structures already existed in the form of ed and grep and you are literally using unix tools... goddamnit I need to learn more about unix now, this is just brilliant!
Fantastic. I've been using and administrating Unix since the mid-1980s back at Bellcore. It's always a pleasure to see one of the early developers talk about the history.
OMG, I remember being a AT&T technician that visited Bell Labs in Holmdel, NJ (I think) in about 1974 or so. We were a new hardware maintenance group for DEC PDP-11s and were given a demo of Unix including a demo of GREP. We were blown away with how GREP piped to other utilities like WC could do really useful analytics. Was still using it when I retired as an Oracle DBA in 2020 when the pandemic hit. Still occasionally play with it on Ubuntu.
The original Unix filesystem was such a wonderfully elegant piece of software that was the foundation so much else.
I love learning about stories like this! So many brilliant minds in the early era of computing! And here I am trying to align things correctly in a web page.
Love these videos with Prof. Kernighan. I sit in front of a Unix machine of one kind or another practically all day, every day. It's unique and wonderful to hear Prof. Kernighan's insights and thoughts on what forms the basis of the devices I use in every part of my life. Thank you for making these Unix history videos, and I am always happy to see more.
Short answer starts at 7:55
(for later reference, you should still watch all of the video)
Thank you.
you da real mvp
Did you find this using a video version of GREP?
g/grep/p
g/re/p
It’s hard for me to stay interested in learning computer science when it’s nothing but code, code, code. But why? Where does this all come from? How did it begin? Why is it important? What did people do before this advancement? This channel answers those questions for me and keeps me interested in learning. Bless you, computerphile ❤️
Your channel should literally be how to teach computers in school.
I love this so much. I thought it was called gREP, not Grep. The history is fantastic! I've used this command so many times compiling software for various reasons.
Thank you.
Nicholas Aranda The computer history museum have a lot of long (and kid of boring unfortunately) interviews with giants of computing
xspager more room for me to attend :-)
Well, since UNIX is case sensitive it is neither gREP nor Grep, but grep. :)
He was amazingly sharp and well spoken for being about 76 at the time of recording. His knowledge and experience just radiates from him.
I just wish Dennis Ritchie were alive today to talk about his work. RIP
I use grep all the time... probably most often to search for running processes via: ps aux|grep or when looking for a file in a directory with a lot of files: ls -lhat|grep ... the uses go on and on though. You can go further in depth and look for files that are open or being used by a program. lsof|grep
This guy is great, I'd like to see more from him.
its Brian Kernigha
He's one of the authors of C
0:39 - M-x mark-whole-buffer RET M-x keep-lines RET RET
I love that he made his notes on dot matrix fanfold paper
More videos with Brian Kernighan please, I really enjoy his talks.
Obviously, all these ED commands are still, in some form, in vi/Vim.
Thank you Ken Thompson for grep!
SED
This kind of thing is invaluable. I use grep multiple times a day, every day and this really ties a piece of software into a real life problem that was solved. I could sit for hours listening to this kind of thing.
Great! Real Expert Perspective!
Generated rich enjoyable product.
Go-on.... reply, Elytron... Please?
Gonna really expect perfection...
Grep real easy patterns
Love these videos featuring Prof. Kernighan! It's first hand, living computer science knowledge and history!
"Non of them were Ken Thompson"
Meanwhile, Ken Thompson with a Wig in the classroom: "This is going to be a peace of cake."
Even being an avid, decades long toolbox command user (like cat, ed, sed, awk, tr, etc. including, ofc, grep) I cannot recall ever hearing this story and it's priceless. Thank you for putting this up.
2:36 "Remember PAPER? Zoom down here, you can see PAPER."
😆
I think he might have originally intended to specify "tractor feed paper", because that is now never used, but commonplace for computers until the late 80s. I remember using Vax terminals in university library in '88. Often people would accidentally print far more than they intended and there was often a long walk between the terminals and the printer, and occasionally you'd see students running through the library to enter a "break" command on their terminal. Meanwhile the feed would pile up in a stack below the printer. Paper recycling was uncommon, so it was pure waste.
I love learning things like this. It’s not only a great history lesson, but is so useful in understanding the functionality of the command itself. Thank you. 👏👏
Where the name "grep" came from? Why, from the verb "grepping through text" of course! .... Wait, what's that? That wasn't a word before? How _did_ people communicate back in the day?
Get a grep on yourself. It's nothing to grep about!
Same with Google. Nobody says "search", everyone says "Google" yourself. Before Google, everyone was finding. xD
I hope anyone that watched was able to get to Seattle last year to see the UNIX@50 display at the Living Computers Museum... it was a dream come true and these videos complement it very well. Much thx!
these history lessons are some of your most important work. its an insight into computer culture you'll never read on Wikipedia or see in a Hollywood movie. its very cool
Absolute genius !!! It's a privilege to watch such people talking about linux and Unix
"none of them were Ken Thompson", lol, so true.
most interesting video I watched over xmas, and the most powerful program ever written. I don't know how I've survive without it.
Kernighan-Lin helped me a lot at work. Thank you!
What about Fiduccia-Mattheyses?
I use this tool every single day and was unaware of the story behind it. Thank you Mr. Kernighan!
Very nice and historically valuable video.
ed still exists, by the way. There's even a GNU version, and its direct descendant vi has a thousand implementations, some even with the ex part. So anyone can try out g//p.
Holy ****, that IS Brian Kernighan! Talk about eating some humble pie. Totally not worthy
Amazing video. Describes how things slowly but surely evolved into the current super computing Era. Not to forget the gigantic efforts of Unix & C pioneers like Ken Thompson & Dennis Ritchie
Raise your hand if you use 'grep' in your everyday speech.
"Grep the closet for your shoes, kids."
"Grepped everywhere for my car keys, can't find them."
Grep is probably the greatest command in my opinion and it was invented overnight. Amazing insight , Ken Thompson is a true programming genius and now he's developing the GO language
Don't forget Rob Pike.
grep wasn't really invented overnight. There is an excellent "fireside chat" between Ken Thompson and Brian Kernighan on this site where Ken is giving some details. Basically, Ken had grep lying around when department head Doug McIlroy asked for a program with such capabilities. Ken gave himself overnight to clean up the program and then delivered, hence it seemed like an overnight development. Ken is still a wizard though and I highly recommend the interview, it is the one from VCF East in 2019. It's a gem!
One of my fav all-time cartoons is the one that compares a group of cavemen with some computer programmers. Both groups exclaim 'grep , awk and mkdir'. It might have been in Kernighan's book. (edited bad spellcheck substitution)
For anyone interested grep coded in C in the original UNIX consisted of two files: grep.c (135 NCSL) and regexp.h (371 NCSL)
Getting ken himself on this channel would be great
+1
Too late
They’d need a time machine or a ouija board
@@TheDavo10001 Buddy, Ken is alive and well. Well, not sure if well, he works at Google, but certainly alive. It was drc we sadly lost a few years ago.
@@gnualmafuerte You mean dmr instead of drc? Yes, sad. I met them both personally and they were nice guys in their own peculiar ways.
Loved this video because it gave me an extra dimension of meaning and understanding to something so common and trivial as the program grep. Amazing explanation, btw!
Very interesting, I always just assumed it stood for "Get Regular Expression Pattern".
Explanations like this are valuable for understanding the context of the things students are learning in modern cs.
oh, grep means: Globally search the lines to find out that match a regular expression pattern and print all of them;
Grep is one of the best tools I ever learned back in my university days and I have been using it for about 28 years. I even use native binary ports on Windows when I use that platform.
You now HAVE TO do the origins of regular expressions
If you haven't done so already, you need to record and make a "history of..." series of videos because your knowledge on these matters and subjects is invaluable to anybody/everybody who uses it.
(As a sidenote, my family and I now are deeply engrossed in Hamilton: The Musical and so my 6-year-old son is starting to learn about The Federalist Papers and it is interesting how that intersected with computer science which gave birth to g/re/p.)
That's such a cool story and I'm so glad that you were able to share it.
grep saved my bacon countless times.
| grep bacon > save
grep, awk, and sed are the holy Trinity of Unix.
Rem Lesk: yes
jqbtube perl != unix
@Markus Glanzer baCRON?
I wish I could convey my teenage son's face when I told someone else about how Bell Labs bought a printer that turned out to not be able to do what they wanted, so they got a guy to hack it. My son had been caught in horrible programming/webbuilding/tinkering addiction through me, but I had no idea how deep he had burrowed on his own. Anyway, I saw on his face that he found it hard to believe that they could decode the innards of a 1970s laser printer... and then I told them that the guy was named Ken Thompson, and my son's face immediately lit up with glee and absolute faith. :)
How did the students do on the assignment?
It's literally just writing a new main() and deleting every function you end up not using...
Fair enough, I'm not familiar with the source of grep tbh
Interesting and entertaining video. Finally I know where GREP comes from. I'll never forget that.
ed is the standard text editor
How appropriate, you fight like a cow
Emphatic. Eloquent. Brian Kernighan. 💕
I remember a parody of “The Sounds of Silence” with the immortal lines:
Sew on this handle that I might schlep you/ Change your permissions that I might grep you...
"Change your permissions that I might grep you" hmm.. where there any privilege escalation bugs in grep? :)
This is an awesome story. Grep is still an amazing tool I use almost every day.
oh come on - don't leave us hanging... did any of the students manage it ?
If any one of them had he would have mentioned it.
Considering it was a computer science task in the early 90's I'd assume everyone taking the class were able to do it. grep isn't a complicated program by any means
The next week's challenge was probably to write the ed script to automatically change the 1800 lines of ed into the grep program. That'd be pretty evil...
He said they were given the ed source code and earlier he said ed included a regex search. So they already had a regex parser. The rest shouldn't be too hard.
+Karl Kastor Indeed if anything it's a lot of stripping down what is there since grep gets rid of all the editing stuff and just focuses on the printing part. Thus of course how sed ended up coming about providing the full functionality of ed but of course with the ability to work on streamed or pipelined input the latter one in particular making it still a staple even today for text manipulation in shell scripts.
Wow, wow, wow! I have used grep for so long and I haven't even once thought about its name. It's just grep, it greps, and that's all to it. Amazing story, thank you!
wow grep was written overnight!?
In Assembler even. I am baffled how that is even possible. Guess you have to be Ken Thompson for that :D
To be fair, if you know what you're going for, it a fairly simple program transformation: remove features and then loop over files. Of course, 70's-era assembly may not have made that easy...
lol yea assembly code sounds like a mountain of 0 and 1's :D
Nope--that's machine code. ;)
Unix was written in a month - a week for the kernel, a week for the filesystem, a week for the shell (but the original Unix shell was not _also_ a _language_ as it is today - that came later with the PWB shell), and a fourth week for init, echo, cat, ed, as, and so forth.
It didn't have pipes, diff, certainly not troff, and fork-exec wasn't invented yet.
Oh my. Rewind. I remember such pleasure of tying a long cmd line with several 'greps' piped together finding the exact data I needed.
Since watching this video my power limit has increased by 25%
For anyone more recently created, what he says at 02:47 may be hard to visualise. And when (shortly afterwards) he refers to "multiple" lines, they amounted in practice to 6, or if you were really lucky, 12 lines (but such capable monitors had to be "fought over"). I didn't and don't love such aspects of those times - but I now far better appreciate how comfortable, supportive and companiable were my work environments back then.
"ed was the standard text editor"
IT STILL IS.
Felipe Cotti By definition, of course. Says so right in the man page...
Menachem Salomon by definition, by conviction, and by heart.
vi
Thank you for creating C language and Unix. The progress of your creations are great. That time was the hard, very limited computing power.
So I'm guessing sed came from ed? Would like to see a comment or video about this.
ed; editor, sed; stream editor nuf said
And awk was inspired by ed or sed
What an honor getting to listen to some of the greatest influences in computer science. I still have his original book he did with Dennis Richie - _The C Programming Language_. Looking at the editor ed you can certainly see where vi was spawned from as well.
9:44 :)
One of the best videos recently!
wow, g/re/p
I don't recall having watched a video with professor Kernighan, he's a great storyteller
5:06 flipping everyone off
I remember fixing a 80mb disk drive on a pdp11 at Rutherford labs - after loading the diags into the paper tape reader you had to input a sequence of switch combinations to load the registers to boot from the paper tape. The diags would locate bad blocks and make them inaccessible
64k was not that shabby back in the day!
As a newer coder that used Linux as his primary workstation, I only really grep STDOUT, otherwise I use the IDE (which of course also uses REGEX). Super cool to know the history though, and I cannot imagine how helpful this was initially.
I always thought it came from "to grab something" .grab, grep same thing but with a tech twist
I love the way he teaches. The ED commands remind me of the syntax in AWK. Bet there is some link, get Ken knows.