Closed ellerychan closed 8 years ago
This is a pretty largely-scoped issue; perhaps it should be broken down.
Or broken up. Go for it.
Created:
Part 3 will continue here.
I just pushed some initial code for this. I'm starting to build the database schema. I really could use some Jaron input.
I made a schema diagram. I put the .graphml
file in the transition folder. I should somehow attach a web-viewable version of it to the wiki. @jawaad-ahmad in particular, but anyone who knows can answer: What is the best way to keep these two views of the same document organized and synchronized. When I save the .graphml
file I can also save a .png
or .svg
or .pdf
version.
.png
(or whatever) somewhere in the wiki?Also, I abandoned the issue_74_start_django_project branch and started a new one called start_django_project_issue4. I am going to delete issue_74_start_django_project.
Should we settle on a naming convention for issue branches, or does it not matter in Git? I was thinking of something simple like issue74 and that's it.
Should we open pull requests to review documents such as your .graphml
file in the transition
folder?
Regarding keeping them in-sync, should the contributor manually run ImageMagick's convert
to save as the desired format(s)? Should we add a commit hook into the repository to run convert
automatically? Does this mean everyone needs to have ImageMagick installed to do this?
For any generated files, I would recommend some sort of generated
subfolder in which to place any generated artifacts. This can apply to any generated source files as well.
My personal preference is to avoid committing generated files in general where possible and instead incorporate the generation into the build process. Generated files can still be output to the generated
subfolder or some other location.
Sometimes that's not possible or convenient, so in those cases I would be OK committing generated artifacts to the repo.
I understand we can link using [text](absolute-or-relative-path)
. I've started using this to refer to wiki pages from issues. To link to our Process page, I put in a link to [Process](../wiki/Process)
here because we're in the same repo. If I was entering this link into an issue in the brata
repo, I would enter [Process](../../brata.masterserver/wiki/Process)
.
If no one has any objections from wiki pages to source repository files, maybe we can keep the generated .svg
file local to its source and then link into the source repo from the wiki page. Maybe something like [Image](../path/to/file.svg)
?
Should the contributor manually run ImageMagick's convert to save as the desired format(s)?
I don't think that will work for .graphml files. But it is easy to export from yEd into an image format. I suppose that possibly yEd has a command line way to do that, but I don't think so.
If no one has any objections from wiki pages to source repository files, maybe we can keep the generated .svg file local to its source and then link into the source repo from the wiki page. Maybe something like Image?
I ran into two problems:
.svg
to display in the wiki page, so I ended up making a .png
, just so I could move on.If you figure out how to do either of those, I'd like to know, and you might add them to the Process page.
I gave it a try too. No luck.
I started a Wiki section on the Process page.
Should we settle on a naming convention for issue branches, or does it not matter in Git? I was thinking of something simple like issue74 and that's it.
Git enforces no convention. I would prefer to have at least a clue in the branch name indicating its purpose, so I don't have to go chase down the issue to figure out what it is for. Typically the branch name would describe the feature being implemented. Seems good to include the issue number at the end. Not worth making extra rules about it, IMHO.
Anything left to do for this, or can we mark this as resolved?
Superseded by many more specific Issues, so this one can be closed.
Continuation of Simplify and Enhance MS Design brata-hsdc/brata#74 Part 3 only:
Simplify the MS design and operation as follows:
Use a Python-based framework such as Django. Using Python would have made it easier for more people to help, and using Django would provide much of the infrastructure needed to make a REST-style, MySQL-backed, web application.