Trademarks List should be improved

[dave2wave]

https://www.apache.org/foundation/marks/list/
https://github.com/apache/www-site/blob/main/content/foundation/marks/list/index.md

This page includes a large section of registered and a large section of unregistered trademarks that were copied in as html from another location.

There is duplicate id for Lucene-core. Is this file up to date?

With the new pelican site we have some options to process trademark data files and build this page as a template.

We will take this discussion to trademarks@pache.org

[dave2wave]

Sent an email to trademarks@ https://lists.apache.org/thread.html/r75233dcc0e18b4015b96f766463ae5b15ca0aff938163e12da5d0755%40%3Ctrademarks.apache.org%3E

[markt-asf]

I haven't been doing a very good job of maintaining this data. It is largely as Shane left it when he handed over to me. There are some CSV files that I believe are the source for most, if not all of the data, on that page in the private foundation/Brand svn repo.

If you were able to convert them to a suitable format I can get an updated list of marks from counsel and between us it shouldn't take too long to get the data updated as we don't register that many marks over time.

[ShaneCurcuru]

My process was to periodically update the list

• Get docket spreadsheet from counsel
• Remove/massage columns for just what we need
• Add rows from projects.a.o for any project names that counsel is not tracking (i.e. for all non-registered/applied for names)
• Use a cgi on whimsy to convert CSV to JSON
• Checkin the JSON

The list.cgi can both do the conversion, as well as reading and displaying the json to be the pretty page.
https://github.com/apache/whimsy/blob/master/www/brand/list.cgi#L58

[dave2wave]

If www-site has the JSON checked in then asfdata.py plugin can convert this to sequences which ezt can convert to html and/or markdown.

[dave2wave]

I tried: https://whimsy.apache.org/brand/list.cgi?json

{
  "exception": "#<Errno::ENOENT: No such file or directory @ rb_sysopen - /srv/svn/brandlist/docket.csv>",
  "backtrace": [
    "/usr/lib/ruby/2.7.0/csv.rb:641:in `initialize'",
    "/usr/lib/ruby/2.7.0/csv.rb:641:in `open'",
    "/usr/lib/ruby/2.7.0/csv.rb:641:in `open'",
    "/usr/lib/ruby/2.7.0/csv.rb:715:in `read'",
    "/x1/srv/whimsy/www/brand/list.cgi:35:in `csv2json'",
    "/x1/srv/whimsy/www/brand/list.cgi:60:in `block in <main>'"
  ]
}

[dave2wave]

I see docket.json in the private Brand folder. Are there any concerns with a file like that being in public?

[dave2wave]

BTW - This file will be easy to work with and we can remove some of the more sensitive fields.

[dave2wave]

https://projects.apache.org/json/foundation/projects.json will be new as well, but I see no issue with the feasibility of converting the ruby into a combination of python and ezt into html. It will be tedious, but should only take a day or two at most of time. It can be done on a preview/trademarks branch over weeks and then a PR raised to merge back to main.

[ShaneCurcuru]

When I created the listing, I did all my work locally (by setting up a Whimsy dev environment, allowing me to use the .cgi locally to pull in /srv/svn/brandlist/ which was mapped to my local SVN checkout. I'd then manually upload the json to public SVN as needed.

Up to Mark T how to do this going forward, in terms of what data to include publicly. A key factor is what columns end up coming to him from counsel's official docket Excel sheet (which indeed has additional private columns).

[dave2wave]

The docket json can be served off of whimsy and not checked into GitHub.com. It will work either way.

[markt-asf]

I've requested an updated spreadsheet from counsel

[dave2wave]

Once you receive this then let's review the format and discuss the process.