On Tue, Jan 22, 2019 at 6:32 PM Hector Santos <hsantos=40isdg.net@xxxxxxxxxxxxxx> wrote:
My opinion.
My only concern is the perception that the IETF is now "requiring" to
learn a new suite of 3rd party tools for a single purpose - RFC Draft
submissions publishing.
We already ask people to learn tools! Have you tried working with a newcomer through the arcana of xml2rfc?
On the contrary, when I've been working with newcomer coauthors on a draft in GitHub, and I mention making a new I-D, they're like "Why do we need to make a special version for IETF?"
To Michael's point, this doesn't advocate for GitHub in particular. But clearly wherever possible, we should prefer the tools that contributors know. That's how we get the bottom-up contributions we're supposed to value in this organization. Personally, my experience is that the people who are contributing to the stuff I'm working on know GitHub.
--Richard
For people doing this all the time, and
probably also using the same tools for other parts of their career, I
can understand it would be productive, but not for the occasional author.
After several decades, I believe an application level IETF online RFC
publishing tool should be available. In the past, I used XML2RFC (a
java app) to outline, produce and publish my drafts. Isn't this
available any more? I would think a HTML5 version would be doable
today, and of course, some vcs would be integrated at the backend.
I personally don't want wish to be learning git details and all the
other scripting tools and text formats for a single purpose. I would
if I have to at some top level rudimentary level just to get the job,
but it is not desirable, and certainly not a career requirement for me.
I entered this thread only because I do plan to write a few drafts and
so it was interesting to see the discussion and more people using
GitHub not only for IETF drafts publishing but also for other things
like as a common web site to illustrate ideas and concepts. I rather
not see that become the norm, but it does shows people need sharable
tools and web sites within the IETF for IETF purposes.
--
HLS
On 1/22/2019 5:04 PM, Michael Richardson wrote:
>
> (Wow, what a lot of emails about how everyone should have choice, as long as
> they choose github.)
>
> The reason I posted about sr.ht is that I strongly feel that authors should
> be using some kind of revision control for your xml or markdown files, and
> that they should have some automation for formatting them (whether Richard's
> wonderful Makefile, five line ones, or a shell script). I don't care what
> people use, but it's hard to host CVS, ClearCase, Mercurial, SVN, etc. unless
> you have some routable public address space of your own. Finding places
> that support
>
> There are some people who like git, but don't like the social networking
> aspects and issue tracking portion, and want all of this to happen on our
> mailing lists. There are some for whom the lack of native IPv6 is really
> embarassing. (I use NAT64 and DNS64 to access github...)
>
> I saw sr..ht as an answer for those people. (It doesn't have an IPv6 yet either)
> There are also a bunch of people who use bitbucket, as basically storage.
>
> I have personally been using git and a simple Makefile to manage my drafts
> for more than a decade. I have taught quite a few people this process.
> I'm really glad that this part has taken off.
>
> While I use github an aweful lot as a way to keep track of things with my
> co-authors, I don't particularly like or want to accept new issues that way.
> I *way* prefer to have a cogent email on the WG ML, *followed* by a pull
> request, ideally with the core of the suggestion in the ML.
> While I'm not as old as some, I'm clearly not young anymore.
>
> I see the web interface as often a distraction. Semi-technical
> (i.e. writers with a non-developers background, and young developers who
> never left an IDE) think that github *is* git, and don't take the half a day
> to actually learn to use git. I've been through this *repeatedly* in other
> fora. Having said that, let me repeat that I *do* use github for many
> things. But not all things.
>
> --
> Michael Richardson <mcr+IETF@xxxxxxxxxxxx>, Sandelman Software Works
> -= IPv6 IoT consulting =-
>
