Advertisement
If you have a new account but are having problems posting or verifying your account, please email us on hello@boards.ie for help. Thanks :)
Hello all! Please ensure that you are posting a new thread or question in the appropriate forum. The Feedback forum is overwhelmed with questions that are having to be moved elsewhere. If you need help to verify your account contact hello@boards.ie
Teaching programming - a few thoughts
Options
-
21-09-2009 1:21pm#1
Comments
-
-
-
-
-
-
Advertisement
-
-
-
-
-
-
Advertisement
-
-
-
I wouldn't see teaching newbies to use 'vim' as a particularly productive way to start their programming education.[vim is] arbitrary and confusing, and not discoverable, and learning to use it doesn't further your understanding of programming very much
There's an interesting point buried in there. Vim seems odd to someone who started editing with notepad... but it doesn't seem odd to someone who didn't (like me - I predate notepad[notepad] is very simple, and everyone is already familiar with it
). Vim is actually quite logically arranged if you've never used notepad. It makes two assumptions - (1) that you think of writing and editing as the seperate activities they are; and (2) that you can touchtype. I fell into both those categories, so I found it easy. Someone who's used to having a single mode in their text editor (like in notepad) where all the actual editing is done by special key combinations and menus, probably wouldn't see the reason for having seperate editing and insertion modes and using the main keys on the keyboard. Someone who started off on dumb vt100 terminals wouldn't have difficulty seeing the reason and if they could touchtype, they'd appreciate the choice of keys used (as they're based around touchtyping). Trust me, you spend much time typing on one of these things and you'll appreciate them too:
Note the lack of F-keys and the low number of non-alphanumeric keys in general
On top of which, of course learning an editor won't teach them about programming. Learning a language probably won't teach them much about programming either Both are just tools to do the job. But, like all tools, some are better than others. The fact that vi (and vim) are still about is a testament to their suitability. And given that I've yet to find an editor as capable as vim (I mean, you try running notepad on a command line via ssh to a server in another timezone over a dialup connection...) -- and emacs doesn't count, it's a word processing environment, not a text editor -- I'll stick with my preferred tools. And recommend them.
Besides, if you find original Vi that hard because of your notepad background, try cream...0 -
-
There's an interesting point buried in there. Vim seems odd to someone who started editing with notepad... but it doesn't seem odd to someone who didn't (like me - I predate notepad ). Vim is actually quite logically arranged if you've never used notepad. It makes two assumptions - (1) that you think of writing and editing as the seperate activities they are; and (2) that you can touchtype. I fell into both those categories, so I found it easy. Someone who's used to having a single mode in their text editor (like in notepad) where all the actual editing is done by special key combinations and menus, probably wouldn't see the reason for having seperate editing and insertion modes and using the main keys on the keyboard. Someone who started off on dumb vt100 terminals wouldn't have difficulty seeing the reason and if they could touchtype, they'd appreciate the choice of keys used (as they're based around touchtyping). Trust me, you spend much time typing on one of these things and you'll appreciate them too:
Uh oh.
A vim debate - I was trying to avoid that :-)
But seeing as we are here...
There is a line of argument that says that VIM is just as easy to use as notepad in some objective sense; but its merely that we are so used to using notepad like tools that we think vim is objectively harder, when it is not really.
I disagree strongly with this. I think VIM is harder to learn in a very important way.
I first suggest that all user interfaces have to be evaluated with some reference to the prior experience of the user.
Sometimes the UI metaphors we are familiar with run very deep, to the extent that its not obvious to us we are using past experience to reason about what a UI should do - but that past experience is always present, and forms our expectations. I think its difficult to say that something is in some sense 'objectively' just as easy, but only to people that aren't conditioned to the other UI metaphor.
We might say that Brainfu*k is just as easy to code in as Java - but only to hypothetical Alien species who think in Brainfu*k.
But what's important is how easy it is to learn for the people who need to learn it, and taking the learners background into account.
Most beginning programmers have experience with GUIs, and familiarity with tools like notepad - maybe even notepad itself.
They don't have familiarity with tools like VIM. They've seen wordprocessors, grown up with them, and they don't see writing and editing as separate activities - and this makes VIM hard to use for them, in a very real sense.
I note that perhaps you were speaking broadly about UIs, and not about the new programmers - but let us agree that notepad is going to be less confusing and less distracting to newbies than VIM.
The next thing I'd like to mention about VIM is that VIM is not 'discoverable' in the way that notepad is. And this is a big black mark against the learnability of vim, in my book.
If you imagine someone with little experience of computers confronted with Notepad vs VIM.
The press the keys - in notepad, the corresponding text appears. In VIM, funny command mode things happen.
They want to save their file.
Notepad has several prominent buttons at the top of the screen - if they press these, they see further actions - one of which is 'save'.
Its discoverable - you can go from very little and discover the rest.
VIM is not like that. Unless someone explicitly tells you what the save command is, or you go and look it up, you will not find it by yourself.
The same applies for undo'ing a change, and most of the other features that notepad supports.
So VIM is not as discoverable as notepad. It is not as easy to learn.
VIM is less forgiving for keystrokes in the wrong place.
Commands stack in VIM. Pressing 3, and then pressing d has a different effect than just press d. Yet there is no feedback that the pressing of 3 has changed anything. This is not obvious in the way that the effects of pressing 3 in notepad are obvious. Notepad is more direct manipulation.
Now, obviously VIM is much more powerful and useful than notepad. It may be hard to find how to do things in VIM, but there are a lot more things to do. And that also, VIM is not nearly as hard to learn as many people make out, their views coloured by their prior experience.
I never ever use Notepad, and I use VIM all the time.
But VIM is harder to learn than notepad - both because the target audience already know notepad, and also because VIM is just in any practical sense, inherently less discoverable.
Learning a programming language doesn't teach you all of programming, its true.On top of which, of course learning an editor won't teach them about programming. Learning a language probably won't teach them much about programming either Both are just tools to do the job. But, like all tools, some are better than others.
But I suggest that in any real practical situation, you are going to learn more programming from the act of learning a programming language than you will from using an editor.
While knowing what a for loop does doesn't mean you how to use one, I think the process of learning it will teach you more about programming than that of learning VIM.
And, for newbies starting out, I'd rather their time was taken up with learning the language, and learning to program, rather than learning to use a complex text editor that is a different style than they are used to.
Hmm - not sure I agree.The fact that vi (and vim) are still about is a testament to their suitability.
Like the qwerty keyboard?
I think thats another good example. Takes a while to learn, is non obvious at the start, and once you learned it, you are unlikely to try to learn dvorak.
But that, alone, isn't testament to the suitability of qwerty.
Qwerty, like VIM, is installed on most unix machines when none of the alternatives are - that makes it very useful to know - but that doesn't mean its the best.
Yes, agreed. Also, its syntax highlighting kicks the crap out of notepad.And given that I've yet to find an editor as capable as vim (I mean, you try running notepad on a command line via ssh to a server in another timezone over a dialup connection...)
But again, this is testament to its ubiquity, and the high inertia it possesses, rather than its fitness for purpose, or overall goodness at the job.
Perhaps if we didn't have VIM on every unix machine over ssh, we'd have a better way of streaming sessions to eclipse (or some generic format, or better tool) thats widely deployed, and just as easy to setup, instead? Its impossible to say.
I don't like emacs either :-)-- and emacs doesn't count, it's a word processing environment, not a text editor
I don't find it hard; I think I first used VI over 10 years ago, so I'm not struggling to learn it or anything.-- I'll stick with my preferred tools. And recommend them.
Besides, if you find original Vi that hard because of your notepad background, try cream...
But I still wouldn't recommend it being used in an introductory programming class - it'd just distract people from the stuff they are supposed to be learning.0 -
Well, why not...But seeing as we are here...
The Sapir-Whorf Hypothesis has never been proven...We might say that Brainfu*k is just as easy to code in as Java - but only to hypothetical Alien species who think in Brainfu*k.
Yes, as I said. But that doesn't mean Vim is doing it wrong. Writing and Editing are seperate activities - that's why you have authors and editors in publishing and have had for a very, very long time.they don't see writing and editing as separate activities - and this makes VIM hard to use for them, in a very real sense.
We shouldn't try rewriting hundreds of years of experience based on a bad UI in a microsoft product
Neither is C. Or brain surgery. Or theoretical physics.VIM is not 'discoverable' in the way that notepad is.
Look, some things are hard. And some tools need to be learnt and they're not newbie-friendly. You don't give someone a tool and stand back in a woodworking class, you teach them to use it first, and carefully so they still have fingers at the end of the first day.
Programming's the same. You need to expend effort to learn to use the tools. Even those things cited as being wonderful for neophytes, like python, still require effort to learn. And there's nothing wrong with that and it doesn't act as too much of a mark against something's suitability for use as a teaching tool.
Hmmm. Translate that to, say, surgery...If you imagine someone with little experience of computers confronted with Notepad vs VIM.
The press the keys - in notepad, the corresponding text appears. In VIM, funny command mode things happen.
"I don't get it, I poked him here with the scalpel and he just stopped breathing. It's a lousy user interface and we shouldn't teach students with it."
*cough*gvim*cough*Notepad has several prominent buttons at the top of the screen
True - though the really important things you learn are independent of the language used (so long as that language can do certain basic stuff). That's why Knuth's books are still so fundamental even with an invented assembly, invented computer and pseudo-code.Learning a programming language doesn't teach you all of programming, its true.
But I suggest that in any real practical situation, you are going to learn more programming from the act of learning a programming language than you will from using an editor.
No, but the original design motives of qwerty are sufficient to throw "suitability" out the window[qwerty] Takes a while to learn, is non obvious at the start, and once you learned it, you are unlikely to try to learn dvorak.
But that, alone, isn't testament to the suitability of qwerty. Vim's design motives are a lot better. And the fact that you have such a large range of options with editors, and such ease in switching, but vim still persists isn't the same thing as the hardware longevity of qwerty...
I'm not sure. In some cases, making it look unfamiliar can be of benefit for learning.I still wouldn't recommend it being used in an introductory programming class - it'd just distract people from the stuff they are supposed to be learning.0 -
-
-
The Sapir-Whorf Hypothesis has never been proven...
Well, I'm not sure that's relevant here. The seems to say that how you speak affects how you think. (am not a linguist!)
All I was saying is that its possible to concieve of a species that found brainfu*k intuitive - but the possibility of such a being doesn't mean you can argue on that basis that brainfu*k is just as intuitive as java.
For all practical purposes, and in the only sense that matters, brainfu*k is horrible - its bad.
Whether what language you think it affects how you see the world isn't exactly the same thing.Yes, as I said. But that doesn't mean Vim is doing it wrong. Writing and Editing are seperate activities - that's why you have authors and editors in publishing and have had for a very, very long time.
Well - I don't consider the fact that being an author and being an editor are separate professions as very strong evidence that something like VIM should have a command mode.
As I touched on before, just because that's how something has been done doesn't mean its the best way to do it - could be an accident of history.
But I acknowledge that if the author/editor divide found in publishing was relevant here, that would be at least a good point, even if not totally compelling.
But - I think its a little different when the author and editor are going to be two separate people, one of whom writes a book, and then hands it over for editing.
But it's generally the one user of vim that's going to be both author and editor; and these two tasks are performed almost simultaneously. The one tool has to be good at both - both tasks are done together by the same person.
Indeed, VIM is fine at this (if hard to use). I'm just saying the fact that both tasks are done separately in the writing doesn't mean that they should be done separately in a programmers tool.
For example, I think the Eclipse/Intellij style IDE, with no explicit separation of the writing and editing tasks is the best tool I've seen for coding in.
No, we shouldn't.We shouldn't try rewriting hundreds of years of experience based on a bad UI in a microsoft product
But that's not what I'm saying. I don't think the literary process is directly relevant here, as I just outlined.
I do think you make a valid point here though. If we go too far down the road of trying to keep the UI experience consistent with whatever the user already knows, we will never innovate, and allow better, more efficient UIs to be adopted.
All I'm saying is that there needs to be a balance too. In an introductory programming course, the emphasis, to my mind, has to be about teaching people to be good programmers. This should mean a focus on teaching them the art of programming.
In an ideal world, I agree that we should teach them the best tools for the job too (I might differ about whether this is indeed VIM, though I do think VIM has a place). And, if I had as long as I wanted to teach the class, and the class were not easily confused, and infinitely motivated, perhaps I would start the course with a few weeks on using VIM.
However, in most scenarios in practice, with a one or two semester college course, I still think the best thing to do is start off with a simple editor like notepad, that they are already familiar with, and that won't get in their way - won't increase their mental load.
I'd be worried that making them start in VIM would just add much additional complexity that is on the periphery of what I'd be trying to teach.
I do think it'd be a fair criticism of C that it isn't discoverable.
Neither is C. Or brain surgery. Or theoretical physics.VIM is not 'discoverable' in the way that notepad is.
If we could make another language that was just as good as C, but more discoverable, then that language would be better.
I think discoverability is important. That's why we write easy to read, well commented code. We might say that 'oh well, its going to be hard, just read it and work it out, you are a programmer, that's not an easy job!'. And some people do say this! But we generally agree that the more discoverable code is better.
I think VIM and C are a bit different to brain surgery or TP.
It's a bit silly to criticise the design of TP or the Brain as undiscoverable - after all, we didn't design the relevant systems! But VIM, on the other hand, is fair game, as far as I'm concerned.
Some things are indeed hard.Look, some things are hard. And some tools need to be learnt and they're not newbie-friendly. You don't give someone a tool and stand back in a woodworking class, you teach them to use it first, and carefully so they still have fingers at the end of the first day.
And I would agree a point I think you are making here, which is that perhaps there's an emphasis on making programming easier than it can properly be made to me. That is certainly not the way to produce good programmers - indeed, I think I said as much in my first post on this thread - I'm in favour of the programmers having to sometimes do it the hard way if that makes them better programmers in the end.
Still would start with notepad though, and move to VIM or Eclipse later, after they've got the basics of coding down, and can't be confused by a whole load of new UI concepts.
Again, we didn't design the UI on the human body. I don't think its a fair analogy as such.Programming's the same. You need to expend effort to learn to use the tools. Even those things cited as being wonderful for neophytes, like python, still require effort to learn. And there's nothing wrong with that and it doesn't act as too much of a mark against something's suitability for use as a teaching tool.Hmmm. Translate that to, say, surgery...
"I don't get it, I poked him here with the scalpel and he just stopped breathing. It's a lousy user interface and we shouldn't teach students with it."
If the human body was something we designed, then, yeah, I think it'd be a fair point that its a bit hard to discover what's wrong with it, and that maintenance operations are more error prone than they need to be.
Given that that didn't happen (any creationists on the programming forum?) its not such a fair analogy.
Yeah - gvim is much better from the discoverability point of view - if you told me I had to introduce first years to programming, and I had to use VIM, I'd probably go with gvim, because thats most familiar and discoverable.*cough*gvim*cough*
Yep, absolutely.True - though the really important things you learn are independent of the language used (so long as that language can do certain basic stuff). That's why Knuth's books are still so fundamental even with an invented assembly, invented computer and pseudo-code.
But... I think a version in a clearer language, more familiar to the readership, would make them more useful.
Not so much if you are going to read them cover to cover, and you are willing to commit to this from the start - in that case the sunk costs of learning MIX are less important. But I given that most people won't do that, I think more people would get more from the books if they were in a more familiar language - even C, or gnu as style asm. That's just pragmatism.
I think talking about Knuths use of MIX is a fair analogy for much of the discussion we are having here on VIM.
Its certainly a good point that the original design motives of VIM were to make a powerful editor that was useful for programmers, and not in in qwerty to (allegedly, depending on sources) cripple the thing.No, but the original design motives of qwerty are sufficient to throw "suitability" out the window Vim's design motives are a lot better.
Still, I mainly brought up qwerty to show that just because something was ubiquitous and widely used is not a strong argument that its the best solution to the problem.
Ah, but to use your earlier point, try using the rest of them over ssh on a remote machine you can't install software on...And the fact that you have such a large range of options with editors, and such ease in switching,
There's definitely a lot of inertia in the continued use of VIM too - maybe that's not as important as the merits of the editor - but I think both play a part.
That's an excellent point. Hadn't considered that - perhaps the ability to start fresh, and discard existing habits, and ways of thinking about things, is more important from a teaching point of view than the initial difficulty of learning the tool.but vim still persists isn't the same thing as the hardware longevity of qwerty...I'm not sure. In some cases, making it look unfamiliar can be of benefit for learning.
Hmm - well I think that's most of the intelligent contributions to the VIM debate that I can make, for better or worse.
Interesting points though - might give Herbal Deity back the can of worms.0 -
-
Advertisement
-
-
-
-
-
-
-
-
Advertisement