5 ms·

I wish we could use Git for scientific papers and legislation. I think the "Pull Request" model makes a lot of sense in both of those cases.

fintler 13y ago
The last paper I published (IEEE for SC12) was written almost exclusively through a private Github repo. We didn't even think about using anything else -- it just worked. It helps that all of the authors were comfortable with TeX however.
- Fomite 13y ago
  Yeah, GitHub for papers is really hampered in fields that don't regularly use TeX.
  - TeMPOraL 13y ago
    In work I accidentally discovered once that Microsoft Word (and Excel) supports three-way-merges natively, so there's at least that for non-TeX folks.
  - Myrmornis 13y ago
    Now that MS Word uses zipped xml as their format you can make some progress by specifying a custom diff driver which pipes the plain text to stdout. https://www.kernel.org/pub/software/scm/git/docs/gitattributes.html https://www.kernel.org/pub/software/scm/git/docs/gitattribut...
  - eru 13y ago
    Just go with markdown?
    - Fomite 13y ago
      As much as I'd like to, I doubt my collaborators would go for it - I work with a lot of clinicians using ancient computers with a stable, predictable work pattern, and they are understandably content with that.
  - hugofirth 13y ago
    People who don't use TeX for accessibility reasons might want to give it another look. The tooling has risen in quality and accessibility hugely of late. I use TexPad[1] - which I like enormously. Otherwise, as another commenter mentioned, there is always Markdown. Combined with MathJax[2] and custom style sheets you see many of the same benefits with greatly increased simplicity. A great desktop Markdown editor (if people don't like web based ones) is Mou[3]. Lastly - the advantage to using Markdown is that if you have to work with someone who does use TeX - something like pandoc[4] can be used to convert between formats fairly nicely. One thing I think is missing is a "Make" like program to dictate how to combine multiple Markdown or TeX documents into a single document, which folders to scan for references etc. I've seen a couple of home grown solutions - but nothing that really fit the bill. [1]: https://www.texpadapp.com/ https://www.texpadapp.com/ [2]: http://www.mathjax.org/ http://www.mathjax.org/ [3]: http://mouapp.com/ http://mouapp.com/ [4]: http://johnmacfarlane.net/pandoc/ http://johnmacfarlane.net/pandoc/
hugofirth 13y ago
We do use Git for scientific papers internally. Especially in the proof reading stage we find it is the only manageable way to have multiple people provide constructive feedback to a paper in a limited time frame. This has caught on with almost everyone (technically minded I should add) who has been exposed to the approach, so I wouldn't be surprised if this started to catch on in the wider field of Scientific publishing and Conference calls.

fintler 13y ago

The last paper I published (IEEE for SC12) was written almost exclusively through a private Github repo. We didn't even think about using anything else -- it just worked. It helps that all of the authors were comfortable with TeX however.

Fomite 13y ago
Yeah, GitHub for papers is really hampered in fields that don't regularly use TeX.
- TeMPOraL 13y ago
  In work I accidentally discovered once that Microsoft Word (and Excel) supports three-way-merges natively, so there's at least that for non-TeX folks.
- Myrmornis 13y ago
  Now that MS Word uses zipped xml as their format you can make some progress by specifying a custom diff driver which pipes the plain text to stdout. https://www.kernel.org/pub/software/scm/git/docs/gitattributes.html https://www.kernel.org/pub/software/scm/git/docs/gitattribut...
- eru 13y ago
  Just go with markdown?
  - Fomite 13y ago
    As much as I'd like to, I doubt my collaborators would go for it - I work with a lot of clinicians using ancient computers with a stable, predictable work pattern, and they are understandably content with that.
- hugofirth 13y ago
  People who don't use TeX for accessibility reasons might want to give it another look. The tooling has risen in quality and accessibility hugely of late. I use TexPad[1] - which I like enormously. Otherwise, as another commenter mentioned, there is always Markdown. Combined with MathJax[2] and custom style sheets you see many of the same benefits with greatly increased simplicity. A great desktop Markdown editor (if people don't like web based ones) is Mou[3]. Lastly - the advantage to using Markdown is that if you have to work with someone who does use TeX - something like pandoc[4] can be used to convert between formats fairly nicely. One thing I think is missing is a "Make" like program to dictate how to combine multiple Markdown or TeX documents into a single document, which folders to scan for references etc. I've seen a couple of home grown solutions - but nothing that really fit the bill. [1]: https://www.texpadapp.com/ https://www.texpadapp.com/ [2]: http://www.mathjax.org/ http://www.mathjax.org/ [3]: http://mouapp.com/ http://mouapp.com/ [4]: http://johnmacfarlane.net/pandoc/ http://johnmacfarlane.net/pandoc/

hugofirth 13y ago

We do use Git for scientific papers internally. Especially in the proof reading stage we find it is the only manageable way to have multiple people provide constructive feedback to a paper in a limited time frame. This has caught on with almost everyone (technically minded I should add) who has been exposed to the approach, so I wouldn't be surprised if this started to catch on in the wider field of Scientific publishing and Conference calls.