[R] opinion - sharing problem data and other stuff.

Jeff Newmiller jdnewmil at dcn.davis.CA.us
Thu Aug 14 00:14:15 CEST 2014


Hear, hear. Getting the questioner to slow down and look at what they really want is key, because if they don't know what they want then their question will be unclear, and I hate answering a question that was never really what the OP was looking for in the first place.

One problem I have encountered is questions where key information is in a linked web page that goes away, rendering the posting archive useless. I suppose as long as gists don't "age out" they can be okay for some questions, but I am still way less likely to go digging there to answer questions in the first place.
---------------------------------------------------------------------------
Jeff Newmiller                        The     .....       .....  Go Live...
DCN:<jdnewmil at dcn.davis.ca.us>        Basics: ##.#.       ##.#.  Live Go...
                                      Live:   OO#.. Dead: OO#..  Playing
Research Engineer (Solar/Batteries            O.O#.       #.O#.  with
/Software/Embedded Controllers)               .OO#.       .OO#.  rocks...1k
--------------------------------------------------------------------------- 
Sent from my phone. Please excuse my brevity.

On August 13, 2014 2:50:16 PM PDT, Sarah Goslee <sarah.goslee at gmail.com> wrote:
>Hi John,
>
>People do sometimes link to external data and code, though I'm not
>sure I've seen any in that particular format.
>
>But, two things to consider.
>
>A. I'm lazy. If the problem is fully-formed in the email, I'm more
>likely to try to solve it than if I have to go download something and
>figure out what's in it. (Even leaving aside the potential issues with
>downloading random things.)
>
>B. A *small* reproducible example is usually a good thing. There are
>not that many cases where the whole big dataset is necessary. Further,
>the exercise of creating a small reproducible example is often enough
>to solve the problem, without ever needed to bother R-help at all.
>
>Granted, I'm also likely to skip posts to the list with enormous
>dput() data dumps too. I'm a big fan of the "small" part.
>
>Sarah
>
>On Wed, Aug 13, 2014 at 5:41 PM, John McKown
><john.archie.mckown at gmail.com> wrote:
>> This is just a thought that has occurred to me. I don't know if it is
>> an "Oh, WOW!" or an "Are you KIDDING?!?" type thought. So I thought
>> I'd ask here.
>>
>> I use github for a few things. Nothing great, but maybe nice. Anyway,
>> one feature of GitHub is the GIST feature. What I am used to github
>> being for is a project consisting of many complete files. A gist can
>> contain many files, but is really for a set of snippets. Such as code
>> sequences. Or maybe the output from a dput().
>>
>> If I have a "big" problem where I think that having all the data and
>> my attempted solutions available, I think it would be far kinder of
>me
>> to put a _good_ synopsis of the problem here on the list. And a
>> clickable URL to the gist I have created for the problem. That would
>> decrease the bandwidth on the email server. And save space on it.
>And,
>> lets face it, in many cases only a few people are going to really
>look
>> at any given problem "in depth", so why have a huge email go out to
>> the entire community?
>>
>> My idea may not be useful, I really am not sure. But my motive is
>> trying to keep everybody's inbox from overflowing. And make it easier
>> to supply really good data, but only to those who are interested.
>>
>> Thanks for any feedback.
>>
>>



More information about the R-help mailing list