4 ms·

How Git shows the patriarchal nature of the software industry

5 points by davidbarker 11y ago

HelloNurse 11y ago
Very sad, considering that reasonable people with less talent for making themselves unhappy can simply start using another name and email, accept that their old commits are associated with their old name, and recognize that as far as DVCS features go integrity trumps flexibility and changing names is such a bad idea that being reminded of bad times is a necessary sacrifice.
greenyoda 11y ago
Making Git user names immutable only seems to show short-sightedness in the design of Git. Whether this design decision was driven by patriarchal biases is debatable - it was most likely just a quick and dirty solution in a program that Torvalds never imagined would become as popular as it has. (With a little extra work, Git could have mapped internal immutable numerical user IDs to external names, as in Unix's /etc/passwd file.) And I don't think you can logically extrapolate from one program designed by one person to the beliefs of millions of people in the software industry. There's plenty of software out there (including Linux) that allows people to change their names easily, all of which was written by the same software industry.
- HelloNurse 11y ago
  It's not "a little extra work", etc/passwd is far simpler because it isn't distributed. How would you propagate user aliases between repositories? Besides, to what repositories? And how would you make an user authoritative regarding his alias? Or would you allow every repository to have its own aliases? And... Easily more complex than the main version control functionality. Sorry, the only reasonable technical answer to such a feature request for Git is HELL NO.
T2_t2 11y ago
I don't even know where to start with this.
seiji 11y ago
If you ever want to change your name on your commits, say you have been married, divorced or chosen a name more congruent with your gender identity What? It's a quirk of git and mercurial that author identities based on name+email are part of the history itself. There could be a model where authors have a unique ID maintained independently of the history (every new name+email pair gets a uuid, you can update what your uuid points to somewhere else, etc). Conflating the global "Name <email>" format as a conspiracy against non-white-males is a bit out there. the majority of people who get a name change are going to be married/divorced women. It has probably never crossed the minds of the creators of Git that this could be such a huge problem to a non-cis or non-male person. What? If you wrote a book before getting married do you go back, collect all sold copies, and rewrite your name? open themselves up to discrimination if they are forced to reveal their trans status by disclosing their assigned at birth name. What? Commit author names and addresses are completely arbitrary, unchecked, and answer to no authority. Make it up as you go along. There are a lot of people using GitHub who use fully anonymous names and have no discernible gender or race. As a wise man once said, "show me the code." Another option is to use a handle/pseudonym. This is also fraught with problems. ... this can come across as unprofessional to many current and future employers. What? Using online nicks isn't unprofessional to people who employ programmers. It's normal to be a series of random names online. Being fully anonymous online was actually encouraged before some entitled boarding school kid decided he wanted to rule the world by invading everybody's privacy. This has personally affected me. Just another example of how a seemingly small, normative assumption can have profound negative effects on minorities. Interesting, but this is more against a policy of ever naming anything and against ever keeping any records for anything ever. Life sucks, the past sucks, we can't always rewrite our past, but we can outgrow it and become better and stronger and faster in the future to suppress, outgrow, and make irrelevant prior pains. But, as far as Git goes, it does have a way of accumulating same same but different name/email pairs from the history and treating them as one unified identity: http://git-scm.com/docs/git-shortlog#_mapping_authors http://git-scm.com/docs/git-shortlog#_mapping_authors — example: https://github.com/docker/docker/blob/master/.mailmap https://github.com/docker/docker/blob/master/.mailmap
dang 11y ago
There's a substantive point here, which is that software designs based on immutable data can conflict with human needs. Sometimes there are legitimate reasons for wanting to alter records of the past. Append-only logs have the same problem. This is interesting because from a technical point of view immutability is almost always a good property to have, and software designs are increasingly organized around it.
voidr 11y ago
If I change my name today, my emails that I sent yesterday won't magically update for everyone to display my new name. For some reason my birth certificate and all text ever written with my name won't automatically update. Clearly this was a ploy by the evil white mail crowed against the feminists, LGBT and minorities. ...or the author should get a hobby. What if I change my name/email address and I do want my old work to stay under my old identity. I might work for Foo Corporation and then move to Acme Corp. thereby changing affiliation and email address. Being able to better manage identities in a git repo would be a great feature, however in my opinion it's a shameful move to call the git developers "patriarchal", it is pure slander which I think has no place in a technical community. The author should either write a patch for git that implements what she wants, or write a new VCS altogether.

5 points by davidbarker 11y ago

Very sad, considering that reasonable people with less talent for making themselves unhappy can simply start using another name and email, accept that their old commits are associated with their old name, and recognize that as far as DVCS features go integrity trumps flexibility and changing names is such a bad idea that being reminded of bad times is a necessary sacrifice.

greenyoda 11y ago

Making Git user names immutable only seems to show short-sightedness in the design of Git. Whether this design decision was driven by patriarchal biases is debatable - it was most likely just a quick and dirty solution in a program that Torvalds never imagined would become as popular as it has. (With a little extra work, Git could have mapped internal immutable numerical user IDs to external names, as in Unix's /etc/passwd file.) And I don't think you can logically extrapolate from one program designed by one person to the beliefs of millions of people in the software industry. There's plenty of software out there (including Linux) that allows people to change their names easily, all of which was written by the same software industry.

HelloNurse 11y ago
It's not "a little extra work", etc/passwd is far simpler because it isn't distributed. How would you propagate user aliases between repositories? Besides, to what repositories? And how would you make an user authoritative regarding his alias? Or would you allow every repository to have its own aliases? And... Easily more complex than the main version control functionality. Sorry, the only reasonable technical answer to such a feature request for Git is HELL NO.

T2_t2 11y ago

I don't even know where to start with this.

seiji 11y ago

If you ever want to change your name on your commits, say you have been married, divorced or chosen a name more congruent with your gender identity What? It's a quirk of git and mercurial that author identities based on name+email are part of the history itself. There could be a model where authors have a unique ID maintained independently of the history (every new name+email pair gets a uuid, you can update what your uuid points to somewhere else, etc). Conflating the global "Name <email>" format as a conspiracy against non-white-males is a bit out there. the majority of people who get a name change are going to be married/divorced women. It has probably never crossed the minds of the creators of Git that this could be such a huge problem to a non-cis or non-male person. What? If you wrote a book before getting married do you go back, collect all sold copies, and rewrite your name? open themselves up to discrimination if they are forced to reveal their trans status by disclosing their assigned at birth name. What? Commit author names and addresses are completely arbitrary, unchecked, and answer to no authority. Make it up as you go along. There are a lot of people using GitHub who use fully anonymous names and have no discernible gender or race. As a wise man once said, "show me the code." Another option is to use a handle/pseudonym. This is also fraught with problems. ... this can come across as unprofessional to many current and future employers. What? Using online nicks isn't unprofessional to people who employ programmers. It's normal to be a series of random names online. Being fully anonymous online was actually encouraged before some entitled boarding school kid decided he wanted to rule the world by invading everybody's privacy. This has personally affected me. Just another example of how a seemingly small, normative assumption can have profound negative effects on minorities. Interesting, but this is more against a policy of ever naming anything and against ever keeping any records for anything ever. Life sucks, the past sucks, we can't always rewrite our past, but we can outgrow it and become better and stronger and faster in the future to suppress, outgrow, and make irrelevant prior pains. But, as far as Git goes, it does have a way of accumulating same same but different name/email pairs from the history and treating them as one unified identity: http://git-scm.com/docs/git-shortlog#_mapping_authors http://git-scm.com/docs/git-shortlog#_mapping_authors — example: https://github.com/docker/docker/blob/master/.mailmap https://github.com/docker/docker/blob/master/.mailmap

dang 11y ago

There's a substantive point here, which is that software designs based on immutable data can conflict with human needs. Sometimes there are legitimate reasons for wanting to alter records of the past. Append-only logs have the same problem. This is interesting because from a technical point of view immutability is almost always a good property to have, and software designs are increasingly organized around it.

voidr 11y ago

If I change my name today, my emails that I sent yesterday won't magically update for everyone to display my new name. For some reason my birth certificate and all text ever written with my name won't automatically update. Clearly this was a ploy by the evil white mail crowed against the feminists, LGBT and minorities. ...or the author should get a hobby. What if I change my name/email address and I do want my old work to stay under my old identity. I might work for Foo Corporation and then move to Acme Corp. thereby changing affiliation and email address. Being able to better manage identities in a git repo would be a great feature, however in my opinion it's a shameful move to call the git developers "patriarchal", it is pure slander which I think has no place in a technical community. The author should either write a patch for git that implements what she wants, or write a new VCS altogether.