Markdown: Data or stfu



  • Re: @blakeyrat point about the actual value of markdown (often repeated, this is just one of many examples):

    @blakeyrat said in What markup format do you prefer?:

    @dkf said in What markup format do you prefer?:

    … that most people don't care about.

    Again an assertion about Markdown and "most people" with zero evidence backing it.

    Prove what you just typed. Prove it with data.

    "We're all science-y computer science people writing code on science-y science computers."
    "How about using science to prove your markup language is better than the alternatives?"
    "No! Make shit up! Asspull! Write completely unsubstantiated bullshit in web forums!"

    He has a point. I did an initial and half assed search, and I can't find any studies that prove markdown is superior or more efficient in any way.

    I'm of the mind that markdown is shit poop, but in the same side of the coin, I can't prove it with data-based evidence.

    As with any extraordinary claim (markdown is the best), it requires extraordinary proof. But I'm also willing to admit that the opposite could use some proof, too.

    So... forum brain people... how do we do that?

    First question would be...is there any existing data or studies we could look at first? Any one good at the Googles, or who has access to academia-level stuff?

    Failing that, let's say we were to design our own study. What would we test, and how, to meet the standard and rigor needed to conclusively say if markdown is "shit" or "the shit"?


  • Impossible Mission Players - A

    @lorne-kates said in Markdown: Data or stfu:

    What would we test, and how, to meet the standard and rigor needed to conclusively say if markdown is "shit" or "the shit"?

    That is a chore in and of itself. For example, comparing features against, say, HTML, *down is overwhelmingly lacking.



  • @lorne-kates I'm not going to provide any data, but I will provide some context:

    Here is the original "spec" for Markdown:

    It's riddled with inconsistencies and vagueness, and the only real thing anyone had to test their implementation against was a Perl implementation that didn't match up with the "spec" in some cases.

    Markdown was posed as a "safer" alternative to HTML that you could give users access to. Now, pretty much every place where you can use Markdown on the internet is designed for tech-savvy people, so HTML wouldn't confuse anyone any more than Markdown's syntax would.

    Plus, Markdown allows raw HTML inline, so any "safety" gained by restricting the markup language is immediately lost, although you can "regain" the safety by sanitizing HTML. Which means in order to make Markdown safer than HTML, you need to write an HTML sanitizer, which would also work on... HTML.



  • @lorne-kates First of all, please find me a usability study for any text entry format. I'll wait.

    Ok, now that we've gotten that out of our system (I didn't actually check, tbh), there are essentially two paths forward:

    • An experiment
    • An archival study

    An experiment would mean that we find a sample of users, and get them to write a text in a number of different formats. We will then do a post-study questionnaire, to measure each subject's satisfaction with using their assigned entry format. We can also measure the "correctness" of their input, the time it took to complete the text, and their satisfaction with the end result.

    We can do this. In fact, I am logging this as a potential bachelor's thesis.

    We could also instrument a discussion forum such that different users get different means of text entry, and then use runtime logging to collect equivalent data. We would probably want this type of runtime logging even with the controlled experiment.

    An archival study would mean that we look at the evidence already produced. We may, for example, find a couple of different fora with similar topics and user bases, and for example measure the number of entries. The problem would be to define and argue for why the fora are indeed equivalent, and why we would thus expect a similar level of activity. Unless we collect data from so many fora that we can just do a factor analysis and see if the activity correlates with the entry format.

    We may also count the number of invectives near texts about markdown, compared to other entry formats.

    Or, we may take an easier route, and just look at how many of the formatting codes in markdown are also supported in word processors, and how many feature requests there are in the issue trackers for word processors where the equivalent shortcuts aren't supported. Mind you, this route will not tell us the technical expertise of the users of those features, nor their frequency of use, in the word processors.



  • Spent five minutes looking for anything about the usability of Markdown. Found this:

    (J. Rubart, "Synchronous Collaborative Text Editing in Wikis", publication status and year unknown):

    Students, who did not know the markdown syntax well,
    needed a lot of time to search and write according to the
    syntax. For them, the synchronous collaboration helped to
    learn the language.

    (note the unmeasurable weasel words "a lot of time")

    Also found this, but didn't have time to read beyond the abstract:

    http://dl.acm.org/citation.cfm?id=1057043



  • @ben_lubar said in Markdown: Data or stfu:

    Markdown was posed as a "safer" alternative to HTML that you could give users access to. Now, pretty much every place where you can use Markdown on the internet is designed for tech-savvy people, so HTML wouldn't confuse anyone any more than Markdown's syntax would.

    From that standpoint, there’s this gem in the spec:

    The idea is not to create a syntax that makes it easier to insert HTML tags. In my opinion, HTML tags are already easy to insert.

    Quickly followed by:

    Markdown’s formatting syntax only addresses issues that can be conveyed in plain text.

    So … just like HTML, then, but less extensively?


  • I survived the hour long Uno hand

    When I did my course in HCI, they taught us that we should typically assign a task for the user to complete, like Mikael's hypothetical text writing task. The closer to real-world task the better, so maybe you take 2-3 users and put them into a forum and instruct them to have a discussion, or play a forum game of some kind. One useful measure when conducting a study like this is the number of errors made, or the number of times they had to backtrack to correct the formatting. Another is the number of clicks it took them to get to their formatting. In this case, you may want to compare how many times users went to the visual toolbar vs typed out the symbols -- if they even learned how to type the symbols at all. You'd want to compare HTML with BBCode with Markdown. And yes, you might want to do a questionnaire afterward to measure their satisfaction with the input system.



  • Doesn't nodebb has plugins for wysiwyg markdown and etc?



  • @mikael_svahnberg said in Markdown: Data or stfu:

    (note the unmeasurable weasel words "a lot of time")

    The big question is if that is a longer or shorter amount of time to accomplish same/similar goals with alternatives [various Mark-UP]

    I despise Markdown, yet I have to admit when given to a random person (in a business environment who needs to create something "just above" plain text, it seems to have the widest success rate in achieving baseline acceptable results.



  • @wharrgarbl said in Markdown: Data or stfu:

    and etc

    et cetera means "and the rest", so you just said "and and the rest".



  • @ben_lubar said in Markdown: Data or stfu:

    @wharrgarbl said in Markdown: Data or stfu:

    and etc

    et cetera means "and the rest", so you just said "and and the rest".

    Maybe he was talking Unix.



  • @lorne-kates said in Markdown: Data or stfu:

    Maybe he was talking Unix.

    When would it not have been \etc\etera ??? [and yes, I have seem people drop a directory with that name in \etc being "cute"!]


  • Impossible Mission Players - A

    @ben_lubar
    Filed under: Just Nike It


  • Impossible Mission Players - A

    @thecpuwizard said in Markdown: Data or stfu:

    @lorne-kates said in Markdown: Data or stfu:

    Maybe he was talking Unix.

    When would it not have been \/etc\/etcetera ??? [and yes, I have seem people drop a directory with that name in \etc being "cute"!]

    FTFY


  • Winner of the 2016 Presidential Election

    @mikael_svahnberg according to the embed functionality of nodeBB combined with my browser, you found nothing: 0_1505482035034_338ab6cc-7e66-46de-9e03-30d566b0e96e-image.png

    Filed Under: thanks NodeBB+Opera



  • @izzion said in Markdown: Data or stfu:

    FTFY

    Indeed...


  • Impossible Mission - B

    @ben_lubar said in Markdown: Data or stfu:

    @lorne-kates I'm not going to provide any data, but I will provide some context:

    Here is the original "spec" for Markdown:

    It's riddled with inconsistencies and vagueness, and the only real thing anyone had to test their implementation against was a Perl implementation that didn't match up with the "spec" in some cases.

    This is interesting. That's the same guy who wrote Ronco Spray-On Usability, one of the most widely-known articles criticizing open-source usability. And he's the guy who created Markdown?

    /me sits back and waits for @blakeyrat's head to explode.


  • :belt_onion:

    @ben_lubar said in Markdown: Data or stfu:

    Plus, Markdown allows raw HTML inline, so any "safety" gained by restricting the markup language is immediately lost, although you can "regain" the safety by sanitizing HTML. Which means in order to make Markdown safer than HTML, you need to write an HTML sanitizer, which would also work on... HTML.

    0_1505875296178_drugs.gif



  • @masonwheeler said in Markdown: Data or stfu:

    @ben_lubar said in Markdown: Data or stfu:

    @lorne-kates I'm not going to provide any data, but I will provide some context:

    Here is the original "spec" for Markdown:

    It's riddled with inconsistencies and vagueness, and the only real thing anyone had to test their implementation against was a Perl implementation that didn't match up with the "spec" in some cases.

    This is interesting. That's the same guy who wrote Ronco Spray-On Usability, one of the most widely-known articles criticizing open-source usability. And he's the guy who created Markdown?

    Maybe this quote from that article explains Markdown:

    the foreward for the book


Log in to reply
 

Looks like your connection to What the Daily WTF? was lost, please wait while we try to reconnect.