This site has the potential to be like Stack Overflow but focused on Ada. But Stack Overflow has some shortcomings that we can over come here.
Limitation of a Help Forums
Duplication of questions
Stack Overflow tries to solve this by making it ageist the rules
This site suggest related “topics” based on the title
Similar questions with significant but nuanced differences are labeled as duplicates
Questions are often closed as duplicates on Stack Overflow when they really were different
As a reference, the reader may need to piece together info from multiple related questions to build the answer to their specific need.
Solution
A reference wiki for lessons learned can keep the compilation of answers in an organized comprehensive form. Where examples and explanations are organized by topic and readers can find answers to there question and continue to read for a more complete understanding.
When a question is answer in the forum the answer can be integrated into the existing reference wiki. As more similar questions are asked the info in the reference wiki for that topic can be refined.
When asking a question, the reference wiki could be used to simplify the question and focus in on the part missing from the wiki.
When answering a question, the reference wiki could be updated to cover what was missing and reference in the response.
Why not Wiki Books?
While Wiki Books’ Ada Programming is a good resource, it may not be easy to integrate into the forum. And the content we might want to add may not fit well into the goals and guidelines of that site.
Hi,
the process seems pretty resource intensive, unless there is an hoppe to get somme LLM doing the job of completing the wiki from forum answers, and filtering questions from already answered parts.
(An interesting project, BTW).
If not, our community has already a hard time maintaining some useful web page, like the wiki book, and is polluted by too many obsolete but still in line site, not sure creating another one is realistic.
However, I certainly don’t want to discourage a good initiative!
(Sorry if this reply is badly edited; I’ve worked on it too long and need to focus on my paying job.)
I agree that this would be useful. Ada has its own particular jargon, and idioms, that trip people up all the time. Something akin to an Ada FAQ that does things like
Translate modern jargon to Ada jargon.
Common idioms and use cases.
I have this vague memory that something like that exists, though… in fact, can you follow this link and say whether it’s similar to what you’re saying? (Look especially at “How-Tos” and “Advanced Techniques”.) If so, what it really needs is someone willing to transcribe the answers into that area… which I think we can do by submitting a pull request to its GitHub repository. So
Is that correct?
Is that welcomed?
How could we advertise and encourage that more?
I would add that unless I misunderstand you, a book is probably not what someone would be looking for to answer these questions.
That aside, let me speak up in defense of StackOverflow. In my experience, StackOverflow’s biggest limitations are:
Users visit, read, then leave. They rarely up-vote useful questions and answers.
The Ada content is only slightly larger than ε, perhaps because the number of Ada-literate visitors is roughly the same size.
The HTML/CSS/JavaScript content is frequently out of date, and not infrequently wrong, an unfortunate consequence of the fact that Web standards and idioms, browser behavior, and browser hacks change often and frequently.
The Python content is similar to HTML/CSS/JavaScript content, except with Python at least one answer is usually correct for some version of the language… just not, perhaps, the one you’re stuck using.
While duplicate questions are against StackOverflow’s rules, I’d be surprised if their being deleted has been a serious problem. Someone wrongly flagged one of my questions and it was deleted, but that was apparently by me, perhaps because it was down-voted and polluted by idiot comments. However, (a) I can still see the question (you can’t, it turns out – following the link takes you to the question someone claimed I was duplicating), (b) I can apparently undelete it (a bright red “Undelete” link appears when I look at it), and (c) I re-created the question, adding a note explaining why it wasn’t a duplicate, and it remains unmolested. Meanwhile, I routinely come across questions that are duplicates, yet have stayed up for years or even decades.
On the other hand, StackOverflow has over the decades built up an excellent tag-and-search system for software development. I’m not sure any other publicly available source comes close.
Finally,
As a reference, the reader may need to piece together info from multiple related questions to build the answer to their specific need.
This strikes me as a positive. Don’t get me wrong; I’ve cursed StackOverflow (and the software world in general, yes including the ARM) many times for not having the one correct answer to my specific question available immediately, but that’s usually my problem, as the correct answer + low view count of some of my very specific SO questions demonstrate.
@LionelDraghi I agree that it would be more work if it was done as a separate task. But I hope that, if integrated well and made easy to do, improving the Lessons Learned might become part of the dialog process rather than just more work done after the fact.
@cantanima I get it. No worries. And it seem good anyway.
Yes, those would be good to add too.
Yes, something like that would work. But some changes would be needed to achieve what I’m imagining:
Currently, when starting a new topic, there is the help thing that pops up on the side suggesting other topics that may relate to/repeat the topic you are creating. I would like to see the “Lessons Learned”, or what ever we are using, be searched too.
Being able to pull up and navigate the “Lessons Learned” on the side of editing a post/reply would make it easy to link to sections and quote them.
Being able to propose edits and share those in the forum discussion would enable improving the content to be one and the same with answering questions.
Imagine this Scenario
I’m imagining a user with a question starting a new topic.
The user starts creating their post, and the help thing suggests a section of the “Lessons Learned” that might be related to their question. The user reads what seems reliant and finds it answers part of their question but not everything they need. Their post will now be more focused than originally intended because they have already got part of their answer. Instead, the post now focuses on the missing part.
The user can reference the section and quote from it. The user could say something like:
This section says <quote> but is unclear about <such-and-such>. Doesn’t say anything about <something>. I want to do <something> but it is missing this needed info. To help clarify, I am trying to do <something>. How would I do that.
Another user could then respond with clarifying questions if needed, and/or suggest changes/improvements to that section that might address the question and improve the document.
The original poster could then provide feedback on the changes, indicating if his/her question is satisfied by the improvements.
The changes may go through additional peer review before being finalized, but the work of asking and answering the question was integrated with the process of improving the documentation.
Of course, nothing is forcing the users to make use of this capability. It would be optional. But if people start using it, and it provides the users value in the topic thread, then I think it would become commonly used.
I didn’t intend to imply that Stack Overflow is bad. (I’ve used it quite a bit.) I was really just trying to point out how we could improve on it and do things our own way.
If your specific question is something like, how do I weave together these different language aspects to achieve my unique goal, then yes I agree that I would not expect that to exist in a single post/thread.
But if the issue is that a specific language aspect is not fully covered, leaving out details that would be needed in order to figure out how to do your own weave, then I think it is a short coming of the resource.
