Open ben221199 opened 2 years ago
I've done this (independent from this ticket), see: https://www.ietf.org/archive/id/draft-bucksch-autoconfig-00.html
Cool. Can I help somewhere?
The repository can be found here: https://github.com/benbucksch/autoconfig-spec
I just found out that @benbucksch is the author of https://wiki.mozilla.org/Thunderbird:Autoconfiguration:ConfigFileFormat, which history goes back to atleast 2008. Sorry that I wasn't are of that. Woops. 😬 Cool that there is a draft for standardization available: https://datatracker.ietf.org/doc/draft-bucksch-autoconfig/. It is a nice starting point for standardization and a good alternative for autodiscover from Microsoft.
For standardization, I think it is important to have a clear understanding about the version and version history of the format. For example, the <clientConfig>
has the attribute version
with value 1.1
in it's latest version. However, 3.1
is also used in older versions and 3.0
is mentioned too. If more clarified what happened with those numbers, it it easier to write a specification, also for older clients.
I crawled to some wiki page changes and got the following versions:
Without version
attribute, but with xmlns
attribute.
3.0
1.0
?With version
attribute containing 3.1
3.1
version code originates from the Thunderbird version1.1
one, but just before renaming?With version
attribute containing 1.1
1.1
version code does NOT originate from the Thunderbird version3.1
one, but just renamed?If the above information (3.1
== 1.1
, and maybe 3.0
== 1.0
too) is correct, I would suggest to standardize both 1.0
and 1.1
.
Hello @benbucksch and others,
Recently I came across autoconfig because Thunderbird uses it. I think this is very useful protocol to set email easily. However, at the moment the protocol isn't standardized yet.
I would like to see autoconfig standardized. This means that
autoconfig
will be registered at https://www.iana.org/assignments/well-known-uris/well-known-uris.xhtml and the XML format (maybe also the old format) and the used HTTP URLs (maybe also the old paths) will be known by both email client implementers and email server hosters. It also will cause the protocol to be more mainstream and more used.If I can help writing this standard (I think making an RFC is the easiest), I would like to do that.
Thanks in advance