This document is available on the Internet at: http://urbanmainframe.com/folders/blog/20040719/folders/blog/20040719/
I had reason to work directly in the Urban Mainframe's RDBMS this morning while implementing a couple of patches to the CMS.
While I was poking around, I noticed several records in my "user_comments" table that were in the "preview state" (see: The Ultimate Comments Handler).
Of course, this is not unusual. I frequently see "preview" comments appear in the database prior to their publication. However, the comments that I was looking at had timestamps outside of the normal envelope...
“there was a usability issue with the comments handler that was preventing the authors from successfully completing the process”
There were two possible explanations:
I suspected the latter immediately and took another look at the comments handler.
The problem was immediately apparent (once I actually looked for it): the inline preview facility presents unpublished comments almost exactly the same as the published ones. The only real difference between the two is that a preview comment has "PREVIEW: Draft Comment" in comment's header. So users see their comment inline with the others and with the same presentation and assume that the comment has been published.
I realise now that I have to make it much clearer that a comment is unpublished when it is in the "preview" condition. As an interim measure, I have added a note to the bottom of comment previews that explicitly states "your comment has not yet been published".
I will adjust the comments handler in the next day or two to colour-code previews, or to provide some other visual cue. That should hopefully prevent this problem from recurring.
After I'd put the interim fix in place, I logged-on to the database to publish the rogue comments. Comments that are in a preview state have a "preview" column set to "1" in the "user_comments" table, so a simple "update" command, changing the "1" to "0", would add them to their respective threads. However, I realised that it would be unethical of me to publish these comments without the approval of their authors.
Maybe I was wrong in assuming that a usability issue had prevented their publication. Maybe the authors had simply discarded their comments. I couldn't be sure, so I left them as they were.
Ben and Smitty, if you read this, let me know whether or not you want your unpublished comments to go live.