I’m under the impression that many technical folks must be annoyed by the emails they get from Stack Exchange. The messages contain a cute fluffy HTML version and a plaintext cripple auto-generated from the HTML version.
The HTML version is used by most web clients and by any clients configured to use it preferentially. The plaintext version is usually preferred by folks who participate in mailing lists, because it’s the greatest common denominator for all the participants, it’s easier to process automatically, it’s safer, it has a convention for marking quotes, it’s free from visual clutter unless the message is totally broken, … HTML, on the other hand, is used as a strong indicator of spam, e.g. by the LKML. Now, it’s probably clear that I prefer plaintext email.
All sorts of problems arise from generating plaintext from HTML:
- Notification emails contain lots of whitespace and empty lines in plaintext version
- URLs break in plain text emails
- Plaintext part of notification mails does not link to comments and is otherwise broken
- Plaintext notification emails do not link to questions
The plaintext version was introduced to stop SpamAssassin default rules to block the messages and because there had been some demand.
Are plaintext emails used enough to warrant the developers’ work needed to produce their content directly from the same sources as the HTML version uses? … in contrast to extracting their contents from the HTML version. After all, they are just two different serializations of the same data.