Ethical issues in modifying and changing the source code of ado-files and SSC archive packages

[Abdullah Algarni]

Dear Statalisters,

Sometimes, I have modifications or enhancements in published ado-files and packages on the SSC archive. Is it ethical to change the source code, republish it, and acknowledge the author of the primary package?

[Jared Greathouse]

It's only unethical if you modify it such that it isn't original/new.

If somebody wanted to republish scul as scul2 (which I wrote) but they do constrained optimization and add in lots of other new stuff that I didn't do, then I don't care. If they just add a new graphics option or do something that isn't really new, then no you can't cite this as yours since you barely made any real modifications to it.

So, originality and citations are key. For me, so long as they cite me as the original author, and they make real changes to it, then nothing is unethical about this.

[Abdullah Algarni]

Thank you Jared Greathouse!

So u think it's unethical to make some MINOR modification even if I have cited the original package?

For example

*! Originally written by Jared Greathouse and modified by Abdullah Algarni at 15 October 2023.

Maybe one possible solution is to modify the package and keep it unpublished in my PERSONAL directory.

[Jared Greathouse]

I mean, are you publishing it also on SSC/elsewhere?


If it's just for submission to a journal, then there's likely no need to do this since you're not "publishing" the command itself like it's yours.

[Abdullah Algarni]

Thank you Jared 😊

[Richard Williams]

If it is a minor tweak you can suggest to the authors that they make the change themselves.

I've copied code from other programs. When doing so I always acknowledge the source. I also ask the authors if it is ok to copy their code. I don’t think I really need their permission but I think everyone I’ve asked is ok with it.

[Abdullah Algarni]

Thank you Richard Williams, that was really helpful!

[daniel klein]

If you are not going to publish your changes, I think you do not need to do anything more than change the code.

Otherwise, I very much like Richard's approach. Contact the author and tell them about the changes you want to do or have done. If the author agree with your changes then they can publish an updated version of the command. Depending on your contribution, you could be acknowledged in the help file (and/or) code or become a coauthor. I believe this is a better way of improving commands for everyone involved, including users of the command, than having different commands floating around that all do essentially the same thing.

[Richard Williams]

You might also be able to write a program that calls a user-written program. For example, you might be able to use ereturned results to do different calculations or create a new graph or whatever.

the main thing I dislike about this is that the author of the original program might make changes that inadvertently zap your program. With official Stata commands version control usually avoids such problems, but user-written commands are a bit riskier. Also, the other user-written command has to also be installed.

that is why I usually prefer to copy code that I want, with attribution, into my own programs. I can also make whatever customizations I want that way.