This is a big commit, and I apologize in advance for the future
git-blames all pointing to me. This commit does a few things:
1. Merges the website/docs and website/www repo into a single website repo
to be in line with other HashiCorp projects
2. Updates to use middleman-hashicorp
3. Converts less to scss to be in line with other projects
4. Updates page styles to be in line with other projects
5. Optimizes images
6. Prepare for S3 + Fastly deployment with scripts, etc.
7. Removes blog posts (they have been transferred to hashicorp.com with
redirects in place
8. Updated sitemap generation script for better SEO
9. Fixed many broken links
10. Add description to all fields
There are two types of box metadata, both in JSON format, so
we need to minimise the potential for confusion between them.
Renaming the component outside the box file to include the
word "catalog" makes it clear that this JSON document can
catalog potentially multiple boxes in one go.
The metadata is optional whereas the box file is required, so it makes
sense to list the box file first. It's also easier on the reader's
brain to start with the more obvious and easily understandable item
(which they'll probably be expecting to read about anyway) and save
the more surprising item till later.
It doesn't make sense to use the present tense when saying box files are
split into two components, when one of those components is the box file
and one is something else.
One better way of phrasing it would be to use the past tense: "Box files
were split into two separate components" but even that's not completely
correct, because the old format did not include the metadata JSON
document which is one of the new components.
So it's safer to just say that today there are two different components.
This is mostly for cookies and google. If you're logged in to
www, you won't be logged in to apex, and visa versa. Also a Vagrant
Cloud bug but wanted to make it consistent on this end.
Seth Vargo
Adam Spiers
Mitchell Hashimoto
Jack Pearkes