openforcefield / openforcefield.org Goto Github PK

Hmm, the proposed change would hard-code this link to the 0.3.0 toolkit release's SMIRNOFF spec page. This could be a problem because tomorrow we're cutting the 0.4.0 toolkit release, which will make substantial changes to this page (and if we hand-correct that, we'll eventually make more changes requiring an update of this link in the future).

RTD is supposed to have a stable branch which corresponds to the latest tagged release, but that's not building for us. I'll try to figure out what's up with that and get us a perma-link to an up-to-date SMIRNOFF spec page.

I'm going to merge this PR, because an outdated link is marginally better than a completely broken one. But I'm making an Issue to tackle this in the future.

Originally posted by @j-wags in #120 (comment)

Our landing page links to https://openforcefield.org/minutes/cab/, which hasn't been updated since 2020. Maybe we should link directly to the Google doc?

Would there be any security/privacy concerns by doing this? (For the record, I switched that doc from "anyone with the link can edit" to "anyone with the link can view", and gave explicit edit permission to project leadership)

Looks like we need to get Ben Pritchard (@bennybp) onto the members page. Also our Consortium page seems to be out of date as we are at least missing OpenEye. @karmencj anyone else?

The DOI link in this screenshot is broken. I'd submit a PR to fix, but I'm not actually sure where it should point.

Edit: it points to 10.0.51.84/RG.2.2.27587.86562.

The "Contact Us", "About Us" and "Join the Consortium" links don't lead to real pages.

Relevant source is in themes/kube/layouts/partials/footer.html

To resolve this, I propose the following changes:

• Connect "About Us" to link to the "Consortium" page.
• Connect "Join the Consortium" to mailto:john.chodera@choderalab
• Remove "Contact Us"

At some point, the publications summaries have switched from a single sentence to a copy-paste of the entire Abstract, which makes it difficult to scroll through. We should remediate this and replace the Abstracts with a single summary sentence.

There are many clicks between the homepage and installation instructions for any packages (say, the toolkit)

Here's what I see as the minimum path: openforcefield.org → software → scroll to OpenFF toolkit → GitHub → scroll all the way down to "Detailed installation instructions can be found here."

It may be better to have shortcuts to installation instructions on https://openforcefield.org/software/, something like below; even if that is just a box that points to the installation instructions in the docs, it shortens the path

This the line where the URL appears:

https://github.com/openforcefield/openforcefield.org/blob/v2/content/force-fields/force-fields/_index.md?plain=1#L42

Having dates in titles is no longer necessary and they should be removed from all titles in Science Updates posts. This was used for the meeting summaries and to make it easier to find relevant entries on the previous website version.

https://www.netlify.com/features/

Netlify is highly recommended by @dgasmith.

If we run into many issues with Travis deployment for the website, we should consider moving to Netlify.

We have a bunch of talks archived on Zenodo, and we could post those online under the Science page!

Here's the first few to add:

• Tobias crystal modeling work now up on Zenodo and citeable, as of approximately now: https://zenodo.org/records/10038276

We will have to add some code to generate these entries, but it should be straightforward.

Related to #114, #121 (etc.)

There are several other places that have the broken link (topology/smirnoff.html), the FAQ being one of the more problematic. Here's a search that finds them:

https://github.com/openforcefield/openforcefield.org/search?q=%22topology%2Fsmirnoff%22

Posting as a separate issue because it can fixed without fixing the details of getting a working stable docs build discussed in #121.

We should add Google analytics to the config.toml file.

https://openforcefield.org/community/faq/

If you try to cite https://openforcefield.org with Zotero (or similar) the citation is mangled, at least partly due to misconfigured meta properties.

Here is the citation in Zotero:

The page source looks like:

Note that the meta property for URL is //openforcefield.org/, which obviously doesn't resolve to anything. Also note in the page source:

  <meta name="author" content="newpixcom">

I don't know enough about web development to determine if adjusting the parameters in this file will break other parts of the site:
https://github.com/openforcefield/openforcefield.org/blob/29d41fe3f247bdd9c9bac0d38d54f47bc2324c6b/config.toml

Perhaps related to #310, although not the same -- I think some rearrangement has broken news links again. The link at https://openforcefield.org/about/organization/ to https://openforcefield.org/community/news/news/introducing-the-consortium/ should probably direct to https://openforcefield.org/community/news/general/introducing-the-consortium/ instead.

My student Camila Zanette has defended and moved on, so we should remove her from personnel.

In the Parsley announcement, there is an unfinished sentence:

All driver scripts for generating this parameter set, as well as parameter fitting

This may have been intended to give a link to the driver scripts mentioned.

The TensorFlow Core overview page is a great example of a very helpful landing page:

• Menu at top with useful major links
• Tabs with most immediate information on Overview, Tutorial, Docs, coming features
• Link to upcoming meetings
• Prominent buttons for tutorials and docs
• Illustrative code snippets with buttons to run the full example from your browser

@maxentile : Should we remove you from the Team page?

We could move you to an "External Collaborator" if you prefer!

Seems like with Dotson and Thompson working now, and lack of $$$, we may want to remove the software scientist ad. Or do we leave it on just in case we get more funding 3-4 months from now?

https://openforcefield.org/science/collaborative-projects/

This page ranks highly in Google searches:

but is probably out of date. I have no idea if these are active projects or something that was put together several years ago. They don't seem to fully describe the current scope of collaborative projects; in this case it miscommunicates how much our software supports crystals (it makes no effort to) and whether or not we're working on it (I'm not sure we are). I suggest any of

• Update the page to track current projects and commit to maintaining it (i.e. quarterly updates) in perpetuity
• Tag each project with dates to represent when they were conceived and include current status/progress
• Remove this page altogether

A. Dalke suggested to add RSS feed to our website.

https://gohugo.io/templates/rss/#the-embedded-rss-xml

Probably want to make an easy entry point for people who might want to contribute/get involved.

Any link ever published to one of the openforcefield.org news slugs (e.g. the big Parsley announcement, https://openforcefield.org/news/introducing-openforcefield-1.0/) is now broken because these have migrated to openforcefield.org/community/news/general for reasons that are likely not compelling. We should consider moving this back to openforcefield.org or at least generating URL redirects from any old slugs to their new locations so that we don't lose the history of every link any other website had to openforcefield.org.

The landing page buttons are covered by footer when I scroll to the bottom of the page while using a wide browser window.

Using Firefox 111.0.1 on Mac OS 12.2.1 on a 2020 M1 macbook air

Hi,

in the blog post on the release of Parsley (https://openforcefield.org/news/introducing-openforcefield-1.0/) the regularization penalty weight w_reg is introduced, but does not show up in the equations (see section "Regularized force field optimization with quasi-Newton iterations"). I think it should be in front of the regularization term. Or does it maybe refer to something else from the original ForceBalance work?

Regards,
Tobias

I often want to point someone to the current Consortium members, an explanation of the Consortium, or info on how to join the Consortium, but the way the current website is organized, this is impossible.

Can we either add anchor tags to make it possible to link this information directly, or somehow split off the Consortium information? The current division of information does not make it easy to recruit new partners.

New website looks great - just encountering some missing TeX, if that wasn't already caught. I'm on a slow internet connection so it could be on my end, but I did try 3 different hardware (Firefox on Linux, Chrome on Windows, and an Android phone)

https://openforcefield.org/community/news/news/introducing-openforcefield-1.0/

The link pointing to SMIRNOFF documentation on RTD is now broken. I assume it should be replaced with this one: https://open-forcefield-toolkit.readthedocs.io/en/0.3.0/smirnoff.html.

@j-wags, can you confirm?

Katy's photo should be updated

It looks like the HTML formatting and markdown-style links aren't getting along in the 1.2.0 table.

It's often useful to link directly to a specific publication in the Publications page. If we add anchor links, it will make it possible to do this.

https://gohugo.io/content-management/cross-references/

Things like http://linkedin.com/mkgilson need to be changed to http://linkedin.com/in/mkgilson

On https://openforcefield.org/science/, the link to the NIH proposal is broken:

This was almost certainly my fault---can anyone fix?

@Yoshanuikabundi and I were talking about how to make it possible for writers to figure out which OpenFF papers/zenodo entries to cite. While Zenodo entries are available for distinct software releases, one challenging case is @ChayaSt's torsion interpolation paper, which isn't tied to a codebase, but is the basis for important results.

We think that a central "how to cite" page for all of OpenFF would be useful, and it could live on the openforcefield.org website. It could tie in to the current publication yaml system, and contain at least basic instructions like

(The following links are accurate to the best of Jeff's knowledge on Apr 1 2022)

• How to cite FFs (right now basically everyone should cite the main Parsley paper,), additionally they can use the table in the openff-forcefields readme to get the specific FF's DOI.
• How to cite the OpenFF toolkit (Recommend they find their specific version's DOI, but provide the "auto-latest-version DOI" for lazy folks) and SMIRNOFF paper
• How to cite OpenFF Evaluator (Zenodo DOI) and preprint
• How to cite things like BespokeFit (Zenodo DOI, could be updated with publications)
• How to cite special features
  • "If your work uses Wiberg bond-order based parameter interpolation. please cite this preprint

Alternatives considered

We considered having a logging statement in the code that would state which paper to cite, but that seems like it would just add to our overwhelming terminal output, either be untested or add a lot of complexity to test, be hard to extend for further papers, and would confuse the majority of users.

Another thought is to have each repo's docs contain relevant publications that users should cite, but I don't expect that people would scramble to find all the docs in their conda environments to figure out which dependencies were pulled in and how to cite them. This becomes an especially large maintenance issue if we forget to update individual repos' pages and things become actively wrong.

We should update the members page to reflect the changes in 2020. I can get to this next week if needed.

The attached screenshot is what I'm seeing. I did not add any extra line breaks in the source.

I think style sheet needs to be updated but not sure how.

(Screenshot below)

It dawned on me it can be nontrivial to find people's github IDs to tag them. We should list these on the people pages.

Do we have to do something to the templates to add a spot for it?

Where should we put info on upcoming events? Should we have a dedicated tab?

Also, what about meeting minutes? We're required to post these for Consortium meetings. Link to them from under Consortium?

(Also discussing on Slack, but putting it here for the record.)

It looks like our Members page has accidentally omitted @cbayly13

PRs cause a failing status because they try to build the PR with Travis. We should update this to remove any reference to Travis and convert as much as possible to GHA - There's a prototype of this at https://github.com/openforcefield/openff-sphinx-theme/blob/main/.github/workflows/gh_pages.yml