Feedback after 2 months of use

I have a negative opinion of the change to Discourse and feel this is a regression in many aspects.

General appearance

The style is rather clumsy and gives a cluttered appearance to the various screens.

Home page

For a regular contributor, questions are too far apart, thus there are less questions on the screen than under AskBot.

The good point is the colour coding helping to spot already read or modified questions and the red line delimiting this new session.

But the amount of questions listed is “dynamic”. If I scroll down, new questions are added on the fly, making it difficult to evaluate how far from the beginning I am. Virtually, all questions can be listed on a single screen. The previous “paginated” approach was more useful to select how deep I accepted to go for “recent” questions.

Question page

There is not enough difference between comment and answer.

UI controls are not all positioned where “natural” and available tools seem to depend on privileges (more on that later).

Formatting

Although formatting is still based on MarkDown, many structuring codes are not available such as #, ##, ###, … for headings which allowed to split an answer into hierarchical parts.

Even lists with multi-paragraphs items are difficult to write.

I admit this is not a concern for newcomers but it helps skilled contributors to provide nice-to-read and appealing answers.

Workflow

AskBot was based on Q&A paradigm: one question, several answers with everyone of them possibly commented. This had limitations but contributed (in principle if not in practice) to have a “compact” site.

Discourse was initially developed for forums = chronological conversations. Chronological conversations don’t adapt well to reused questions for two main reasons:

  • conversations run around particular cases and this hinders generalisation
  • to avoid losing the timely thread, replies can’t be edited

This second point is the worst. Apparently, Discourse enforces this and prevents contribution edition after a short period (at least for newcomers who are frequently requested to edit their post to add basic information like LO version or OS name). They can’t because of insufficient privileges and we’re back to a lengthy conversation.

This inherently runs into the risk of famous TL;DR (too long, don’t read).

Thus we have multi conversations in a question: comments added to the question itself to improve its understanding and comments under answers to clarify points. In the end, we’re in TL;DR. And the usefulness/reusability of a topic is impacted.

Even the search function does not seem superior to AskBot. In anyway, results suffer from the same drawbacks as the home page: too much space for one hit and dynamic number of items. In AskBot we only have the question title (yes, some of them were not informative) and here we have more with a tag line and the beginning of question/answer or comment where the term appears but this is not always useful to spot the helpful item.

This wastes too much space and is spoiled again by the dynamic addition to the list preventing to see at a glance how relevant the search key was (selective or not, reported by the list length).

Relevance evaluation

In AskBot, question and answers could be upvoted (required privilege).

In Discourse, two features compete against each other.

Upvote (without downvote!) is very insconpicuously located at top left of question/solution but not enhanced by any label or explanation/suggestion.

A “Like” button is offered among the tools below any contribution. Most newcomers are used to social networks or YouTube and prefer this familiar “Like” button upon the mysterious undocumented upvote feature.

Answers are reordered on number of upvotes, thus displaying the most relevant or appreciated answers at top just below the question. From what I see, “Likes” don’t cause answers to be reordered (nor comments either, but since comments are a “conversation” they must remain chronologically ordered).

This ruins the concept of Q&A where we expect the most relevant answer closest to the question and makes searching for the “good” answer user-unfriendly. Once again, there is a risk of TL;DR.

User trust level

This replaced the karma points. However, the privileges are not adjusted as they should and there are many shortcomings.

A user should always be allowed to edit his post, above all newcomers. Presently, edition is possible only during a short time after submission. When out, you can only comment, reverting to a conversation system.

Also, many tool buttons under a submitted question/answer/comment depend on trust level (as an example the edit button mentioned above).

There is also the question of site moderation/administration. The previous site version was notoriously not administered but this was compensated by community “police”. Provided you had enough karma points, you could do virtually anything.

This is no longer the case. I haven’t yet seen spam, but according to tentative edits on others’ questions, there is no longer the possibility to delete other posts.

To take an example, I recently asked to retag a question, another regular contributor did it and acknowledged the request. To lean off the topic, I deleted my comment but was not able to do the same on the reply comment, thus leaving an “orphan” reply comment which is worse than having done nothing. And I’m supposed to be at top-level trust level!

So either this site is administered by an individual (or several) who commits himself to this very time-consuming task or it is vital to set up mechanisms so that it can be community-regulated as it was in the past.

Summary

If we really want this site not to drift into a messy thread of sorts, we need to review its configuration (if this is possible within Discourse framework).

EDIT 2021-10-11 Yet another issue

Please look at how-do-the-users-here-move-a-comment-to-answers/69145 for another complaint.

3 Likes

I can’t Like or Upvote this post enough. In fact, I can’t Upvote it at all. I’ll add one point: on my desktop I use bookmarks to return to the list English - Writer for faster navigation when I’m finished with a topic. In the old format, I could return to the overview of English topics with my preferred tags, here you have to go out if your way to get there. Together with the other problems stated by @ajlittoz , I regularly wonder why people in the LibO community seem to hate user help user boards so much that they make it as hard as possible for users to discuss LibO. This has been the case from the start and was actually the subject of some discussions on the Apache OpenOffice users forum.

1 Like

My procedure is to right-click on a topic to open it in another tab. I keep the English (all tags because i also contribute to Draw and Math) open in its original tab. I only need to refresh the page from time to time. But this workflow seems to disturb statistics collection because I have a very low reading time!

To be honest with Discourse, there seems to be some beginning of an automatic refresh of the overview page, though this is not perfect because refreshing it from the browser gives different results.

The new version of this website is much more difficult to use and less clear than the old one.

1 Like

Hi, first of all, thanks a lot for your feedback, it is really appreciated and will help us to narrow down what needs to be changed on the site.

I believe this is part of a setting, we should ask Guilhem if it could be modified to a longer period (it seems this setting:

post edit time limit

How do you make your search, do you search within some time frame, or by title?

Discourse consider that a flag should be used instead. And several flags by trusted users will hide the post.

The like button could be disable, would it solve the issue?

Then you have to flag the post and leave a comment that it should be deleted, moderators will delete it. Anytime you see something that should be done and you can’t then, please flag it so moderators could take care of it.

I agree with you that we need to review the settings of the site, we have changed what we have seen necessary but we don’t have the usage of the site that you (and other big contributors) have and your feedback is very valuable and welcomed.
Also have you seen my other post on flag settings?

Cheers

Hi,

Well, you know the reason why we had to move from AskBot to another site. There is no ideal replacement for it, we all know it also. Now we try our best to adapt the site to our needs and help our users. If you have requests to change settings, add plugins or anything else that should be done, feel free to let us know.
Cheers

One of the biggest problems with a board that should serve as a questionnaire, where people ask questions, get answers, then mark the post that constitutes the answer to their question as the answer, is the so-called eternal September. Most new members don’t bother to read the “forum rules”, they just ask a question, get an answer, but quite often never tell that the answer was helpful. On the Dutch language section of the Apache OpenOffice users forum I kind of made a habit of thanking people for taking the trouble to tag a question as solved or just writing that their problem was solved. I even remember one fresh user asking if such a feature couldn’t be implemented. It couldn’t, and anyway, I never saw him/her back again. Most of the Solved tags in the OpenOffice forum are added by moderators. I left the AskBot site for quite a while out of frustration and on my return found that my number of accepted answers had roughly doubled, merely because one member kind of made a living out of accepting answers to questions where the OP didn’t do it. You know who I mean ;).

Don’t mistake me - a forum or board where you can show your appreciation of a post is great - but only if all users use that possibility. If the majority don’t, it’s completely useless and in fact misleading.

since I’ve seen this remark a few times, and it doesn’t really match what I perceive from this site, I take the liberty to mention it here :wink:

Needless to say I couldn’t get 10% of the legacy remarks from this old post, and its adjacent ones, but the defintely underlying interesting question is how the knowledge sharing is currently possible on Discourse, … and how is it (could it be) evaluated ( / monitored).