Best Pull Request Ever



  • Just click that. And support it.



  • I suppose you people would gripe at me if I said I have absolutely no clue what this thread is about. What's a "pull request?"


  • ♿ (Parody)

    @blakeyrat said:

    I suppose you people would gripe at me if I said I have absolutely no clue what this thread is about. What's a "pull request?"

    OK, I'd never heard of this before, but after clicking around it makes sense, and it's kinda funny. Right. We're dealing with a DVCS (namely, git). So random other people can work on their clone of your repo, and then I suppose they can request that you pull their changes into their repo, merging your code with theirs.

    Now, for the joke. If you follow the link to the actual patch that the guy is requesting be merged, it looks like he's just deleting all of the code: "Showing 6 changed files with 0 additions and 1,911 deletions." See, on the first page, here's what the requester said about his changes: "Solved a few compatibility issues. Will always cleanly merge. :)"

    So, kinda funny.


  • @boomzilla said:

    Now, for the joke. If you follow the link to the actual patch that the guy is requesting be merged, it looks like he's just deleting all of the code: "Showing 6 changed files with 0 additions and 1,911 deletions." See, on the first page, here's what the requester said about his changes: "Solved a few compatibility issues. Will always cleanly merge. :)"

    So, kinda funny.
    The problem is that the link in the original post takes you to a page where you don't see the actual patch.  You have to know to click on the last link that says "Much Better" which then takes you to the actual patch.  Once you see it then it is sort of funny, but, a joke doesn't work when you tell the beginning wrong.

     



  • @El_Heffe said:

    The problem is that the link in the original post takes you to a page where you don't see the actual patch.  You have to know to click on the last link that says "Much Better" which then takes you to the actual patch.  Once you see it then it is sort of funny, but, a joke doesn't work when you tell the beginning wrong.

    I still don't get why it says there's 1,911 deletions when there's only 6. Does git honestly always do diffs line-by-line? Even if you delete an entire file? Crazy.


  • ♿ (Parody)

    @blakeyrat said:

    I still don't get why it says there's 1,911 deletions when there's only 6. Does git honestly always do diffs line-by-line? Even if you delete an entire file? Crazy.

    Yes, that's pretty common for SCM, actually. Of course, normally, you get some lines deleted and others added. I think that sort of display makes sense generally, since the output format is consistent regardless of what changed. And it's also correct, since it's telling you about how many lines went away.



  • @boomzilla said:

    I think that sort of display makes sense generally, since the output format is consistent regardless of what changed. And it's also correct, since it's telling you about how many lines went away.

    Well, at the very least the wording is off. Unless you're actually suggesting that guy went through the file, line-by-line, and hit delete on each one (making 1,911 deletions) instead of doing a Select All and hitting delete once per file (making 6 deletions). So no, no it doesn't make sense, but it could be easily fixed by changing the terminology. Not that an open source developer would give half-a-shit about confusing terminology...

    I think well-designed SCM software usually uses the term "changed lines" or somesuch.


  • ♿ (Parody)

    @blakeyrat said:

    @boomzilla said:
    I think that sort of display makes sense generally, since the output format is consistent regardless of what changed. And it's also correct, since it's telling you about how many lines went away.

    Well, at the very least the wording is off. Unless you're actually suggesting that guy went through the file, line-by-line, and hit delete on each one (making 1,911 deletions) instead of doing a Select All and hitting delete once per file (making 6 deletions). So no, no it doesn't make sense, but it could be easily fixed by changing the terminology. Not that an open source developer would give half-a-shit about confusing terminology...

    I think well-designed SCM software usually uses the term "changed lines" or somesuch.

    I think your uninformed pedantic dickweedery is getting the best of you. I say uninformed because from previous threads, I recall you saying that you hadn't used many SCM packages, and mainly had experience with VSS or TFS or whatever.

    Here's what it says: "Showing 6 changed files with 0 additions and 1,911 deletions." What do you suppose an addition or change to a file could mean? I can only really think of 2 reasonable options here. Characters and lines. But lines is much more obvious. I'm not suggesting that he went through the file, but you're looking at the diff of the commit. The software is looking at two different states of the repo, and see that a lot of lines that used to be there, in these 6 specific files, are gone. I suppose you could make the output more verbose, but, meh. It seems like complaining that a form says "First" and "Last" instead of "First Name" "Last Name" or something.



  • @boomzilla said:

    I think your uninformed pedantic dickweedery is getting the best of you. I say uninformed because from previous threads, I recall you saying that you hadn't used many SCM packages, and mainly had experience with VSS or TFS or whatever.

    I've used CVS, SVN, TFS, a few others. I'm not sure how many I have to use before I'm qualified to talk about a text caption that's obviously fucking wrong as any English-speaking person can see in a millisecond. Lemme know if I'm qualified or not.

    @boomzilla said:

    Here's what it says: "Showing 6 changed files with 0 additions and 1,911 deletions."

    Yes, but that's wrong, because there's only 6 deletions.

    @boomzilla said:

    What do you suppose an addition or change to a file could mean? I can only really think of 2 reasonable options here. Characters and lines.

    Seriously? Fucking Unix people, no fucking imagination at all.

    How about undo-able blocks of text? (That is, the amount of text you type that a reasonable Undo algorithm would consider a single block). How about the most fucking obvious answer, the one I've been assuming because it's so fucking obvious, but contiguous blocks of text that differ between the last check-in and this check-in?

    @boomzilla said:

    I'm not suggesting that he went through the file, but you're looking at the diff of the commit. The software is looking at two different states of the repo, and see that a lot of lines that used to be there, in these 6 specific files, are gone.

    Obviously I fucking understand that that is the number its reporting. But that's not the number it says it's reporting. It says it's reporting "deletions", when it's obviously not. (Unless, as I said before, the developer hit the delete key 1,911 times.) If it said it was reporting the number of "deleted lines", it would be correct. If it said it was reporting the number of "changed lines", likewise, it would be correct. Why is this concept so hard? I'm looking at an obvious bug.

    Call me a pedantic dickweed if you want.

    @boomzilla said:

    I suppose you could make the output more verbose, but, meh.

    I don't care about verbose-ness, I care about correctness. Right now it is not correct.

    @boomzilla said:

    It seems like complaining that a form says "First" and "Last" instead of "First Name" "Last Name" or something.

    And I'd complain about that, too. So I don't see your point here.


  • ♿ (Parody)

    @blakeyrat said:

    @boomzilla said:
    I think your uninformed pedantic dickweedery is getting the best of you. I say uninformed because from previous threads, I recall you saying that you hadn't used many SCM packages, and mainly had experience with VSS or TFS or whatever.

    I've used CVS, SVN, TFS, a few others. I'm not sure how many I have to use before I'm qualified to talk about a text caption that's obviously fucking wrong as any English-speaking person can see in a millisecond. Lemme know if I'm qualified or not.

    OK, that's different than the impression I got before, so FTFM.

    @blakeyrat said:

    @boomzilla said:
    Here's what it says: "Showing 6 changed files with 0 additions and 1,911 deletions."

    Yes, but that's wrong, because there's only 6 deletions.

    Yay! Deliberate obtuseness!

    @blakeyrat said:

    @boomzilla said:
    What do you suppose an addition or change to a file could mean? I can only really think of 2 reasonable options here. Characters and lines.

    Seriously? Fucking Unix people, no fucking imagination at all.

    How about undo-able blocks of text? (That is, the amount of text you type that a reasonable Undo algorithm would consider a single block). How about the most fucking obvious answer, the one I've been assuming because it's so fucking obvious, but contiguous blocks of text that differ between the last check-in and this check-in?

    I guess you could do that. Here's why it's a bad idea. Depending on diff algorithms, the answer could be very different. You end up with questions like, "Using version 1.1 it says m deletions, but version 1.2 says n. WTF?" Lines are unambiguous, however, so you provide a more user friendly answer by using lines, even if it's not obvious to you.

    @blakeyrat said:

    @boomzilla said:
    I'm not suggesting that he went through the file, but you're looking at the diff of the commit. The software is looking at two different states of the repo, and see that a lot of lines that used to be there, in these 6 specific files, are gone.

    Obviously I fucking understand that that is the number its reporting. But that's not the number it says it's reporting. It says it's reporting "deletions", when it's obviously not. (Unless, as I said before, the developer hit the delete key 1,911 times.) If it said it was reporting the number of "deleted lines", it would be correct. If it said it was reporting the number of "changed lines", likewise, it would be correct. Why is this concept so hard? I'm looking at an obvious bug.

    The bug is your reading comprehension or something, not with the software.

    @blakeyrat said:

    @boomzilla said:
    I suppose you could make the output more verbose, but, meh.

    I don't care about verbose-ness, I care about correctness. Right now it is not correct.

    Bullshit. Only because your pedantic dickweedery is preventing you from interpreting anything other than your feigned ignorance allows.

    @blakeyrat said:

    @boomzilla said:
    It seems like complaining that a form says "First" and "Last" instead of "First Name" "Last Name" or something.

    And I'd complain about that, too. So I don't see your point here.

    Just that TRWTF is your pedantic dickweedery, so thanks for reinforcing that.



  • Fuck you. You're just trolling me. There's no way you're so stupid you can't see that obvious bug.


  • ♿ (Parody)

    @blakeyrat said:

    Fuck you. You're just trolling me. There's no way you're so stupid you can't see that obvious bug.

    Fuck yourself. I'm convinced that you're trolling me.



  • @boomzilla said:

    @blakeyrat said:
    Fuck you. You're just trolling me. There's no way you're so stupid you can't see that obvious bug.

    Fuck yourself. I'm convinced that you're trolling me.

    no u!!



  • Why are you two arguing about the meaning of deletions when it's a useless concept for the get-go? Automatically outputting statistics from a diff tool is basically a really expensive random number generator.

    Actually it's better than a random number generator, because it's real-life noise instead of a pseudo-random algorithm.


  • ♿ (Parody)

    @dhromed said:

    Why are you two arguing about the meaning of deletions when it's a useless concept for the get-go? Automatically outputting statistics from a diff tool is basically a really expensive random number generator.

    You have a good point. I think I just got drawn in by blakey's ignorance trolling.



  • @boomzilla said:

    You have a good point. I think I just got drawn in by blakey's ignorance trolling.
     

    You should probably propose to him.



  • Is anyone watching the new Beavis and Butthead episodes?  "Daughters Hand" is about a pull request.



  • Just had to think of that, too.



  •  @blakeyrat said:

    @boomzilla said:
    Here's what it says: "Showing 6 changed files with 0 additions and 1,911 deletions."
    Yes, but that's wrong, because there's only 6 deletions.

     Actually, because the text specifies a unit of measure (files) for one metric, and does not specify a different unit of measure (lines) for the other metrics, the confusion is understandable.

     The assumption, when a unit of measure is specified, would be that any other measures would be in the same unit of measure.

    For example, if I told you to go to the store and purchase 10 gallons of milk and 8 of Oil, you wouldn't bring me back 8 Quarts of Oil.  No, you'd get 8 Gallons.



  • @Medezark said:

    @blakeyrat said:
    @boomzilla said:
    Here's what it says: "Showing 6 changed files with 0 additions and 1,911 deletions."
    Yes, but that's wrong, because there's only 6 deletions.
    Actually, because the text specifies a unit of measure (files) for one metric, and does not specify a different unit of measure (lines) for the other metrics, the confusion is understandable.

    I'm not entirely sure whether you're agreeing with me or not, but I would like to point out there is no "confusion", the text is literally and obviously incorrect. Changing "deletions" to "deleted lines" would fix it, assuming the software isn't capable of tracking the actual number of deletions. (Which I assume it's not since, a. it's written by Linux people who rarely keep track of details like that, and b. if it did, this little bit of text would probably already be correct.)



  • @Medezark said:

    The assumption, when a unit of measure is specified, would be that any other measures would be in the same unit of measure.

    For example, if I told you to go to the store and purchase 10 gallons of milk and 8 of Oil, you wouldn't bring me back 8 Quarts of Oil.  No, you'd get 8 Gallons.

     

     

    A wife asks her husband, a software engineer, "Could you please go shopping for me and buy one carton of milk, and if they have eggs, get six!"


    A short time later the husband comes back with six cartons of milk.  The wife asks him, "Why the hell did you buy six cartons of milk?"


    He replied, "They had eggs."

     



  • @da Doctah said:

    @Medezark said:

    The assumption, when a unit of measure is specified, would be that any other measures would be in the same unit of measure.

    For example, if I told you to go to the store and purchase 10 gallons of milk and 8 of Oil, you wouldn't bring me back 8 Quarts of Oil.  No, you'd get 8 Gallons.

     

     

    A wife asks her husband, a software engineer, "Could you please go shopping for me and buy one carton of milk, and if they have eggs, get six!"


    A short time later the husband comes back with six cartons of milk.  The wife asks him, "Why the hell did you buy six cartons of milk?"


    He replied, "They had eggs."

     

    Hey! Something funny in the "funny" thread!


  • @da Doctah said:

    @Medezark said:

    The assumption, when a unit of measure is specified, would be that any other measures would be in the same unit of measure.

    For example, if I told you to go to the store and purchase 10 gallons of milk and 8 of Oil, you wouldn't bring me back 8 Quarts of Oil.  No, you'd get 8 Gallons.

     

     

    A wife asks her husband, a software engineer, "Could you please go shopping for me and buy one carton of milk, and if they have eggs, get six!"


    A short time later the husband comes back with six cartons of milk.  The wife asks him, "Why the hell did you buy six cartons of milk?"


    He replied, "They had eggs."

     

    Shouldn't he have bought seven cartons, though?



  • I'm getting a 404 on that pull request link. I wonder if it might have been spammed by trolls or something? Just a guess.


  • 🚽 Regular

    @blakeyrat said:

    Changing "deletions" to "deleted lines" would fix it, assuming the software isn't capable of tracking the actual number of deletions. (Which I assume it's not since, a. it's written by Linux people who rarely keep track of details like that, and b. if it did, this little bit of text would probably already be correct.)
     

    Wait... from what I gather, you're saying if a user selects all text and hits "delete" that is 1 deletion. Yet if the user goes line by line and hits delete, that's a deletion for each line. And that doesn't account for character deletions.

    My question is, then, wtf is the point of getting statistics on how many times a user has hit the "delete" key? The SCM It's not going to keylog the editor for each time you're deleting something, nor would that be useful to anyone except perhaps for some overzealous sports analysts who might be interested in how many times the delete key was hit in some software studio while Tim Tebow is making a forward pass.

    Personally, I'd probably vouch for a statistic that might look something like this:

    Net Character Count: -553

    Lines Affected: 28

    Files Affected: 3

    Still, in the end, if you're really concerned about what changed in a specific revision, the only really useful thing is to diff the damn file and be done with it.



  • @RHuckster said:

    Wait... from what I gather, you're saying if a user selects all text and hits "delete" that is 1 deletion. Yet if the user goes line by line and hits delete, that's a deletion for each line.

    Yes.

    Out of curiosity, what did you think the word "deletion" when used as a noun meant?

    @RHuckster said:

    And that doesn't account for character deletions.

    If the user selects 5 characters, and hits "delete" (or "backspace", or "Cut" or "Delete" from the Edit menu. or type another character over the selection) then the deletion is 5 characters long. I really have no idea how people are having trouble with this extremely simple concept.

    @RHuckster said:

    My question is, then, wtf is the point of getting statistics on how many times a user has hit the "delete" key?

    I didn't say there was a point to it, or that it was useful information to have. Don't put words in my mouth.

    What I said is that the text claims to be reporting that number, but is not. Which is a bug.

    (Although taking the complexity up a notch, generally you wouldn't consider each press of the "delete" key to be a deletion. For one thing, there are many other ways to delete text from a document, some of which I listed above, and most of which don't make use of the "delete" key at all. For another, if a user presses delete 16 times in a row, you probably want to count that as a single deletion for various different reasons-- the most obvious being the Undo history.)

    @RHuckster said:

    The SCM It's not going to keylog the editor for each time you're deleting something,

    Then it shouldn't claim to be reporting that information. That's all I'm saying. It's very simple.


  • Discourse touched me in a no-no place

    Given the obvious context of the page (and numbers) in question, why are some* people deliberately missing the point of the apparent humour in it? I ask merely out of amusement, since I'd hope everyone else knows the answer.


  • 🚽 Regular

    @blakeyrat said:

    I didn't say there was a point to it, or that it was useful information to have. Don't put words in my mouth.

    By saying Unix/Linux users don't have any imagination for not having such a feature, you sure damn well imply it. Reading the thread, I don't see any reason to not believe you think having this kind of information is a good idea. In fact, you are so adamant about this feature, you're willing to add it to your ever expanding everyone-is-a-troll-or-an-idiot-why-oh-why-doesn't-anyone-think-like-me repitoire. Sure, you could change the wording to fit the description of the information more, but it sounds like you're just as willing to keep the vague wording in, and instead make some useless and even more confusing numbers out of those labels.

    So, excuse me for putting words in your mouth, but I assumed that by spending the time preparing and baking those words for yummy consumption, you'd be willing to eat them. Now, they're going to go to waste and become stale.



  • @PJH said:

    Given the obvious context of the page (and numbers) in question, why are some* people deliberately missing the point of the apparent humour in it? I ask merely out of amusement, since I'd hope everyone else knows the answer.
    http://forums.thedailywtf.com/forums/p/25149/269887.aspx#269887


  • Discourse touched me in a no-no place

    @Sutherlands said:

    @PJH said:

    Given the obvious context of the page (and numbers) in question, why are some* people deliberately missing the point of the apparent humour in it? I ask merely out of amusement, since I'd hope everyone else knows the answer.
    http://forums.thedailywtf.com/forums/p/25149/269887.aspx#269887

    Well quite, but you repeat what I said. You know the answer.



  • @RHuckster said:

    By saying Unix/Linux users don't have any imagination for not having such a feature, you sure damn well imply it.

    No I didn't.

    @RHuckster said:

    Reading the thread, I don't see any reason to not believe you think having this kind of information is a good idea.

    Then your reading comprehension sucks.

    @RHuckster said:

    Sure, you could change the wording to fit the description of the information more, but it sounds like you're just as willing to keep the vague wording in, and instead make some useless and even more confusing numbers out of those labels.

    Then your reading comprehension sucks.

    @RHuckster said:

    So, excuse me for putting words in your mouth,

    No. You owe me 5 Cokes. ICE COLD.


  • 🚽 Regular

    @blakeyrat said:

    No I didn't.

    @blakeyrat said:

    Seriously? Fucking Unix people, no fucking imagination at all.

    How about undo-able blocks of text? (That is, the amount of text you type that a reasonable Undo algorithm would consider a single block). How about the most fucking obvious answer, the one I've been assuming because it's so fucking obvious, but contiguous blocks of text that differ between the last check-in and this check-in?

    Whoops, I'm so sorry. You're right. I said Unix/Linux users. You were referring to Unix users! Omg, my reading comprehension does suck!

     



  • Everyone in this thread above this line is an idiot and wasting my time.



  • @Sutherlands said:

    Everyone in this thread above this line is an idiot and wasting my time.

    Who yanked your chain?



  • @RHuckster said:

    Omg, my reading comprehension does suck!

    I'm more interested in hearing your definition of the word "deletions", when used as a noun in relation to text editing.

    The fact that a supposed programmer, who spends his entire career using text editors, doesn't understand extremely basic concepts of text editors, well, that kind of worries me.



  • @frits said:

    @Sutherlands said:

    Everyone in this thread above this line is an idiot and wasting my time.

    Who yanked your chain?

    1) You have to look at the tags.  2) This thread is stupid


  • @Sutherlands said:

    @frits said:
    @Sutherlands said:
    Everyone in this thread above this line is an idiot and wasting my time.


    Who yanked your chain?
    1) You have to look at the tags.  2) This thread is stupid

    Your name is above the line. ... so is that the joke? ...or was it just an oversight on your part?

    In either case, I want to hear more about this chain.



  • @Xyro said:

    @Sutherlands said:
    @frits said:
    @Sutherlands said:
    Everyone in this thread above this line is an idiot and wasting my time.

     

    Who yanked your chain?
    1) You have to look at the tags.  2) This thread is stupid
    Your name is above the line. ... so is that the joke? ...or was it just an oversight on your part?

    In either case, I want to hear more about this chain.

    Yes, that would be the joke...

     

    But this thread is still stupid.


  • 🚽 Regular

    @blakeyrat said:

    I'm more interested in hearing your definition of the word "deletions", when used as a noun in relation to text editing.

    I'll humor your request and say that, in relation to text editing, a deletion is a character or a set of characters that was deleted by the user. Not that it at all relates to SCM because SCM isn't a text editor. You do know that source control isn't directly related to text editing, right? You can commit binaries, too. When I think of deletions when related to source control, I don't think of edited text, but sets of changes to the types of files at hand. You know, deleting a function, for example, would be a "deletion" in that context. Deleting an entire class would be another deletion, or an image file, or a few characters, or a line of code. If I see readme.txt was changed, then I think of deletions in terms of sections, paragraphs, or sentences.

    Semantically, "deletion" is arbitrary in meaning, and on that matter we agree and Git should make label it more clearly. When SCM has a "deletion" count, it's meaningless, since a deletion could be a line, a file, or a contiguous block of lines or characters. For that reason, though, it's stupid to assume when SCM counts "deletions" it's counting any of those units (or the number of deletions in relation to a text editor). You're just as wrong to assume that as anyone else is wrong to assume it's counting lines, characters, or files.

     



  • @Sutherlands said:

    And why do you want to hear about my "chain"?

    Sorry for the word play.  It won't happen again.



  • @RHuckster said:

    I'll humor your request and say that, in relation to text editing, a
    deletion is a character or a set of characters that was deleted by the
    user.

    So... you agree with my definition, and yet somehow disagree that Git's wording is wrong.

    @RHuckster said:

    When I think of deletions when related to source control, I don't think of edited text, but sets of changes to the types of files at hand. You know, deleting a function, for example, would be a "deletion" in that context. Deleting an entire class would be another deletion, or an image file, or a few characters, or a line of code.

    I don't see how that differs from the above definition.

    I'm ignoring the points about SCM not being only used for text files because:
    1) it's irrelevant
    2) it doesn't change the fact that Git's wording is wrong
    3) I'm a firm believer that SCM should be more closely integrated with your text editor, so you could get some of the great features you can see in, say, Word's revision tracking. But since the industry is dominated by Unix "one application-> one task" people, that'll never happen. And besides that's another conversation altogether.

    @RHuckster said:

    Semantically, "deletion" is arbitrary in meaning,

    No it's not.

    @RHuckster said:

    and on that matter we agree and Git should make label it more clearly.

    It's not an issue of clarity (and thus Blakeyrat begins to fucking repeat himself because nobody reads his fucking posts), it's an issue of CORRECTNESS. The text there now is not correct. Changing it to be something like "lines deleted" would make it correct.

    @RHuckster said:

    You're just as wrong to assume that as anyone else is wrong to assume it's counting lines, characters, or files.

    I assume it's counting deletions, because that's what it says it's doing. Alas, it's lying to us.


  • 🚽 Regular

    @blakeyrat said:

    So... you agree with my definition, and yet somehow disagree that Git's wording is wrong.
     

    Where did I say I disagree with you on that account? Please tell me, since your reading comprehension is apparently superior to mine. I never said "deletions" was a good label to use; to the contrary, I've said I agree with you on that argument. All I said was your interpretation is misguided. See, whenever I come across an arbitrary label, I don't go out on a limb and assume it means something I think it means. If I come across a document that has "Price:" on it, I don't assume it includes taxes, S/H, and other fees, nor do I assume it doesn't. What I do is investigate further and see whether that price includes it or doesn't. That way I don't get surprised when I made the decision to purchase.

    @blakeyrat said:

    I'm ignoring the points about SCM not being only used for text files because:
    1) it's irrelevant

    This whole thread is about the word "deletions" when used in an SCM. So, yeah, it is relevant.

    @blakeyrat said:
    I assume it's counting deletions, because that's what it says it's doing. Alas, it's lying to us.


    Ugh, it's counting deletions as it applies to SOURCE CONTROL MANAGEMENT AND NOT TEXT EDITORS. Source control tracks deletions of files, directories, and file contents, unlike a text editor which only tracks deletions of file contents... so a deletion in SCM is different than a deletion in a text editor, and thus it's asinine to apply deletions as it relates to text editors to deletions as it relates to SCMs. Get. That. Into. Your. Fucking. Head.

  • ♿ (Parody)

    @blakeyrat said:

    3) I'm a firm believer that SCM should be more closely integrated with your text editor, so you could get some of the great features you can see in, say, Word's revision tracking. But since the industry is dominated by Unix "one application-> one task" people, that'll never happen. And besides that's another conversation altogether.

    Since there's no reason to keep one of these threads to one conversation....

    I think most people would say that making the SCM interface richer isn't a bad thing. But the problem is getting all of those features into all of the various tools that people use to edit files. Of course, then you'll have issues where one SCM does something differently, or has a completely different paradigm, and so that doesn't even make sense with the SCM. The actual useful intersection of SCM and editors where this would make sense would likely be so small that it probably already exists, and stays small because it sucks.

    I wonder which track changes features you're actually thinking of?

    There are actually extensions for svn at least that integrate with Word's track changes feature.



  • @RHuckster said:

    See, whenever I come across an arbitrary label, I don't go out on a limb and assume it means something I think it means.

    What, to you, is the point of labeling anything, ever? Then?

    @RHuckster said:

    Ugh, it's counting deletions as it applies to SOURCE CONTROL MANAGEMENT AND NOT TEXT EDITORS.

    You might have missed the part where I said I don't understand what difference you believe there is between the two. Maybe after you explain that, I'll be able to properly reply to this.

    @RHuckster said:

    Get. That. Into. Your. Fucking. Head.

    Maybe if you actually explained your position, I'd get it into my fucking head. As-is, there's nothing to put in my head: you made an assertion I don't agree with/understand, you haven't bothered explaining it. Impasse.



  • @boomzilla said:

    I wonder which track changes features you're actually thinking of?

    Well, at the very least, Word's little color-coded bubbles saying who added what to the file would be handy. Maybe a slider on the top of the window so you could tell it how far back to go-- set the slider all the way to the right, and the bubbles disappear, move is leftward and they pop-in. Further left = further back in time.

    I could also envision a system that would be smart enough to know where the natural language boundries are, so you could put a little left and right arrow on every function/codeblock/whatever and slide it left to go back in time (with the color-coded label showing who made that commit), and right to go back forward.

    I mean, really, there's a thousand possibilities out there absolutely nobody's working on, since the Linux people don't see the value in it, and Microsoft and Apple haven't (for whatever reason) innovated it yet. (The reason probably is: because Linux people are in charge of IDE development at Apple and Microsoft.)

    This is another one of those "if I had $5 million in angel investor money I could dominate this industry in 5 years" things. There are so many of those...


  • ♿ (Parody)

    @blakeyrat said:

    Well, at the very least, Word's little color-coded bubbles saying who added what to the file would be handy. Maybe a slider on the top of the window so you could tell it how far back to go-- set the slider all the way to the right, and the bubbles disappear, move is leftward and they pop-in. Further left = further back in time.

    There's nothing quite as slick as that that I'm aware of, but the basic capability exists. VCS generally have some sort of annotate or blame command that links lines with who and when (the line is the standard code unit, of course). Most commands have an option for specifying a particular point in the repository, so you can pretty easily view earlier or later versions.

    @blakeyrat said:

    I could also envision a system that would be smart enough to know where the natural language boundries are, so you could put a little left and right arrow on every function/codeblock/whatever and slide it left to go back in time (with the color-coded label showing who made that commit), and right to go back forward.

    This sounds like what you had above, but basically combined with code folding.

    @blakeyrat said:

    I mean, really, there's a thousand possibilities out there absolutely nobody's working on, since the Linux people don't see the value in it, and Microsoft and Apple haven't (for whatever reason) innovated it yet. (The reason probably is: because Linux people are in charge of IDE development at Apple and Microsoft.)

    Maybe. I think a lot of these basic features already probably exist in many tools, even if they have a different interface than what you might be thinking of. I think that the typical VCS user probably uses a similar amount of functionality of his VCS as does the typical Word or Excel user.

    @blakeyrat said:

    This is another one of those "if I had $5 million in angel investor money I could dominate this industry in 5 years" things. There are so many of those...

    Uh huh.



  • @boomzilla said:

    There's nothing quite as slick as that that I'm aware of, but the basic capability exists. VCS generally have some sort of annotate or blame command that links lines with who and when (the line is the standard code unit, of course). Most commands have an option for specifying a particular point in the repository, so you can pretty easily view earlier or later versions

    If the GUI doesn't exist, the feature doesn't exist. That's basically saying, "every part of that, except the most important by far, already exists."

    @boomzilla said:

    Maybe. I think a lot of these basic features already probably exist in many tools, even if they have a different interface than what you might be thinking of.

    Yeah, I'm thinking of "good." What they have now is, "awful."


  • ♿ (Parody)

    @blakeyrat said:

    @boomzilla said:
    There's nothing quite as slick as that that I'm aware of, but the basic capability exists. VCS generally have some sort of annotate or blame command that links lines with who and when (the line is the standard code unit, of course). Most commands have an option for specifying a particular point in the repository, so you can pretty easily view earlier or later versions

    If the GUI doesn't exist, the feature doesn't exist. That's basically saying, "every part of that, except the most important by far, already exists."

    Except that's not what I said. Instead of something like a slider, you click on different version numbers to change the display. So, is it less slick and polished than that? Sure. But is it really that different from a user's perspective? Probably not. Personally, I tend to prefer to look at a side by side diff between two versions. Watching something that looks animated may be pretty, but is it really very useful?

    @blakeyrat said:

    @boomzilla said:
    Maybe. I think a lot of these basic features already probably exist in many tools, even if they have a different interface than what you might be thinking of.

    Yeah, I'm thinking of "good." What they have now is, "awful."

    Well, frankly, the features you mentioned here don't even sound that useful. Like I said, I prefer to see a side by side comparison rather than something better suited to a kid's magazine (stare at this picture, then look at the next page, and figure out what's different!) than analyzing code. And there are plenty of tools that do a pretty good job showing that sort of comparison.

    I know, I know, there are TONS of other stupendous features that are missing, even if you can't name them.


  • 🚽 Regular

    @blakeyrat said:

    @RHuckster said:
    See, whenever I come across an arbitrary label, I don't go out on a limb and assume it means something I think it means.

    What, to you, is the point of labeling anything, ever? Then?

    Arbitrary label. Not just label. Arbitrary label. Labels are not pointless when they are concise and meaningful.

    "deletions" is an arbitrary label. It doesn't refer to what is being deleted, whether it's lines, files, or whatever. "deleted lines" is not an arbitrary label.

    "price" is an arbitrary label. It doesn't refer to whether it's a discounted price, a price after taxes or shipping or anything else. "sale price: $X + tax", "original price: $X + tax", "MSRP (incl. tax)", or "final price" is not an arbitrary label.

    Reading is FUNdamental!

    @blakeyrat said:

    You might have missed the part where I said I don't understand what difference you believe there is between the two. Maybe after you explain that, I'll be able to properly reply to this.

    Seriously? You're worried that I don't know the basics of text editors, yet you can't tell the difference between a text editor like Notepad or Eclipse and SCM software like Subversion or CVS?

    I... I don't know how to explain the difference much the same way I can't explain the difference between a port-a-potty and a vibrator since they're so vastly different... but I'll try.

    Let's say you have a website, ok? I'll keep it real simple and just say you used Notepad to edit everything. Yes, it doesn't have all the bells and whistles an IDE has, but it gets the job done. In this example, you use Subversion as your SCM of choice. You have a collection of PHP, JavaScript, HTML, CSS, and image files... maybe a couple of swfs, in a directory. You've used your trusty Notepad to make a change to a JS function, say function foo(bar, baz) { return bar; }. You realize that baz is not used in this function, and decide to remove that parameter. You remove ", baz" from that code and then proceed to remove the second parameter from any calls to that function in other files. Ok? Then you see that one of the files that is in this project isn't being referenced anywhere, so now you use Subversion to delete that file from your work space and flags it for deletion from the remote repository when you commit.

    Now you verify that all tests pass, and you decide to commit your changes to Subversion. Subversion commits the file deletion plus all of the changes to the other files that removed that second parameter. The next person to update from Subversion picks up those changes, sees which files have been modified and deleted, and can even use a diff utility to see exactly what changed in each file.

    Okay, now that I've given you the example above, I want you to swap all bolded words such that Notepad is replaced with Subversion and Subversion is replaced with Notepad, and tell me in all honesty if you still think text editors and SCMs are the same damn thing. Yes, I'm aware that Eclipse has a Subversion and a CVS plugin. That doesn't mean they're the same thing. Just like you can use a vibrator inside a port-a-potty and they're still not the same thing.



  • Look if you're just going to sarcastically shoot-down all my ideas, then don't ask.

    The problem is I don't know what would work and what wouldn't because I haven't done usability prototyping/testing. I haven't done usability prototyping/testing because I'm not working on writing a SCM or IDE. Software isn't one guy in a closet decreeing "this is the best way of doing it" and then implementing, software is the long, hard road to finding the best way of doing it. The reason I think I'd be good at it is nobody (within experimental error) is actually doing usability prototyping/testing for SCM and IDEs right now. Do I know all the answers? No. Anybody who says they do is a liar. What I do know is how to find the answers.

    It's ok to realize that something, even if you don't know what would be better. Because if you know it sucks, you can do the research to find what would be better. The real danger is when people start to think the tools we have are good enough, because then nobody's working to improve them-- and that's where I think that particular segment of the IT industry is right now.


Log in to reply