New web site

[painter1]

The new CF Conventions web site is still under construction. Use this ticket to report any problems with the new site. Here are some problems which I am aware of:

1. CF Conventions document, html version, versions 1.6 and 1.7-draft1:

The web server inserts some extraneous code early in Chapter 9, which messes up the rest of the chapter, forward to some point in the appendices. The html document itself is correct.

2. Documents page, both links to v26 of the Standard Names table are dead.

3. The quick links just give you stubs for standard names and the conformance document.

4. The top bar just gives you a stub for "Conformance". It's worth reconsidering, btw, what really belongs in the top bar.

5. There is no link to the document source code (in DocBook?).

6. The system for downloading a pdf document is un-intuitive.

7. The old url cf-pcmdi.llnl.gov/* should be set to redirect to something which works; the new web site if possible.

[ngalbraith]

Just adding a link to the new site, currently the best URL is http://cf-convention.github.io/index.html

[russ]

What still needs to be done to get the canonical URL, ​http://www.cfconventions.org , to link to the new site? Currently, it results in a 404 response from a github server.

[graybeal]

In the spirit of simplicity, may I suggest the best current URL can be ​http://cf-convention.github.io (the /index.html is assumed by the server), and the canonical URL can be ​http://cfconventions.org (fewer characters).

[graybeal]

Copying from email: I suggest that rather than reporting all problems on one ticket, we create a new ticket for each problem and use the component 'cf-trac-site' or 'cf-web-site' as appropriate. Here's why:

• each problem can be tracked, commented, managed (prioritized, assigned), and signed off separately
• otherwise, this one ticket will never be signed off (because there will always be open sub-problems)
• problems can go in the main ticket Description, rather than in hard-to-find comments scattered throughout
• fixed vs non-fixed problems can be tracked using Trac, including by component
• the ticket name covers the exact problem, and the component covers the topic, making it easy to find and manage problems in multiple ways

[graybeal]

Problem (1) above seems to be fixed. I'm not seeing any issues in either http://cf-convention.github.io/1.6.html or http://cf-convention.github.io/1.7.html, or in the corresponding PDFs.

[caron]

​http://cfconventions.org/standard-names.html

shows page with links that dont have anything to do with standard names, so its effectively blank.

[caron]

​http://cfconventions.org/latest.html

"Lastest CF Conventions Document"

should be "Latest"

[graybeal]

And 1.7 should not be the latest, or rather, latest should really be latest released. Should point to 1.6 (could put a pointer to 1.7 as well, making clear it is the unreleased next version).

[mgschultz]

Just a quick one: is it still necessary to point out so prominently that "The CF conventions generalize and extend the COARDS conventions."? In my view we left those behind now and the new web page could get rid of this ballast. It's of course OK to make this reference within the conventions document, but on the home page???

[jonathan]

Replying to mgschultz:

Just a quick one: is it still necessary to point out so prominently that "The CF conventions generalize and extend the COARDS conventions."? In my view we left those behind now and the new web page could get rid of this ballast. It's of course OK to make this reference within the conventions document, but on the home page???

I agree with Martin, it's now too much of a historical reference to mention on the web page. When CF started, we had to establish its credentials by linking it to other metadata schemes.

Jonathan

[graybeal]

A mild disagreement on the idea of eliminating the COARDS relationship from this page. I've been writing a FAQ that talks about CF conventions relations to COARDS. As I understand it, these points go against CF's current basis: CF does not stand alone, it depends literally and directly on the COARDS standard, from everything I remember reading.

Maybe I just remember too much old documentation, so if we've really gone through the pieces and COARDS has been fully assimilated to the extent necessary, then we should update all the CF materials, not just that page. But if CF is still not complete without the COARDS text, I think omitting that relationship is confusing rather than helpful.

(I spent hours going through the COARDS convention earlier this year to fully understand CF, because CF repeats little of the COARDS information, at least about attributes. This CF weakness is worth fixing; the COARDS standard is oddly presented, and a bit difficult to read and integrate into the CF framework. I just think the front page should match the perspective elsewhere.)

[ngalbraith]

I think CF is fully documented, and users don't need to refer to COARDS documentation to understand it.

If you look at the list of conventions on the ​NetCDF Conventions page, COARDS is described this way: 1995 standard that CF Conventions extends and generalizes.

So it seems to me that CF documentation might mention COARDS as a historical point of interest, but it's not a critical piece of information and does not belong on the (otherwise succinct) CF ​home page.

[robert.muetzelfeldt]

I'm trying to track down the "Guidelines for Construction of CF Standard Names". The old link (​http://cf-pcmdi.llnl.gov/documents/cf-standard-names/guidelines) takes you to a page which links to the new site. On the new site, the Quick Link to "CF Standard Name Table" takes you to an almost-empty page (as mentioned above). Going to Documents and clicking on "Standard Name Table v27 HTML" takes you to a page which does provide a link to the Guidelines - and it's the same one as above...

I guess this is part of the "Under construction" process for the new site, but just thought I'd mention it in case it's been missed.

Apologies if I have missed something obvious or previously explained.

[ngalbraith]

Thanks for bringing this up.

There's a copy of the 2008 version of the guidelines document at ​https://web.archive.org/web/20130728212039/http://cf-pcmdi.llnl.gov/documents/cf-standard-names/guidelines but I'm not sure if that's the most recent version.

It does seem that this should be available in the documents section of the site.

[ngalbraith]

I was looking for this trac page because I noticed that when I access the html version of the current standard name table, the title is 'Program for Climate Model Diagnosis and Intercomparison' - hard to find in the sea of tabs in my browser. This is in the html head/title, not sure how that's generated on the CF site now.

While trying to find this trac site from the cfconventions.org page, I ran into several 404s; I'll go back when I have a minute and see if I can find them again.

[jonathan]

I am closing this ticket because as far as I know the new website is complete and functional (and has been some time). Thanks to Jeff, Matthew and others who migrated it.