Thinking about Xenforo search capabilities

[E-HP]

Some random thoughts....Looking at the original thread regarding the troubleshooting subforum, and rating the solutions, I'm wondering if Xenforo provides the capability of searching on, sorting of, or filtering on, the ratings flags through the search UI ? More broadly, are there any fields, besides things like subject title or member , implemented or not, that can be used to narrow search results?

Virtually everything I've learned about ebikes has been through reading posts, old and new, on this forum. I believe everything you need to know is here, among the many posts. Some posts are really good, or have a really good discussions where all of the info in the thread is helpful and accurate. A 10 out of 10. Some threads have good nuggets of info, but may be polluted with some bad info; maybe a 7 out of 10. Some threads are so bad or provide no future value, that they'd be better off in the trash bin; 0 out of 10. A thread with no responses could probably be auto-flagged as a 0 after a couple of years.

The great news is that all the information is here. The bad news is that it takes tribal knowledge to sift through it and identify the good from the bad. At some point, as with the troubleshooting section, we might want to use that tribal knowledge not just for addressing current problem solving, but taking a moment to rate the thread at or near it's conclusion, in a way recording that tribal knowledge in field that can be leveraged later for members to separate out the bad from the good. We're probably far away from that with the current search capabilities, but even absent that for now, it might be something to think about.

It seems like it's taking a while to get the knowledgebase humming, so I was thinking about things that might be less onerous and take less of a commitment but still moving in the direction creating ways to navigate through this information jungle. It might even help for adding info to the KB at some point, to identify and allow copying and pasting of some good info already in ES, rather than rewriting all of it.

 

[amberwolf]

whenever someone is about to navigate away from the thread they're reading, ask them to rate it. (technology already annoys us with this to ask for subcriptions to various websites when you try to close it, etc., so...)

 

[neptronix]

Xenforo's search engine doesn't provide any of that. We don't have post ratings other than emojis. 'brain' and 'science' emojis could be used in the future to identify good posts. But we have to write a system from scratch i think, to search by emojis.

Xenforo's search engine is quite bad, and the developer team there does not respond to user requests. So we can't expect any help from them.

The best we have currently is the advanced search option, and the google search option.


As for our knowledgebase, it's been an enormous flop so far. A lot of people wanted the feature, so fundraising went well enough. And we built the most convenient to use software we could fit into that budget in hopes that more people would want to write articles. What i didn't bet on is that we still couldn't get anyone to contribute.

Simultaneously, it's the only way we can expect to curate the golden info - sticky threads still don't work.

No feedback has been provided on why we get no contribution. One problem we know of is that the WYSIWYG editor is in some instances finnicky, but equivalent to the forum's post editor in ease of use, though it has more features.
Here's the pending change list for that software that we'd like to get done in the next 6 months:
- In-system notifications - important, as it models how Xenforo works.
- Much improved WYSIWYG editor, but waiting on Xenforo's implementation of it so we can copy it and thus have real 1:1 functionality between the forum and KB ( this is 6 months late )
- Improve user interface and replace icons with ones that are much more obvious

That's our last shot to get people to use it.

Re: your idea of copying and pasting from posts, the question is, where does it go? right now you can copy an entire thread or post into the knowledgebase, and you can expect a minor amount of editing needed to correct the format. This can be done already. You can mark an article as draft so that other people don't see it, then copy it to a new article.

I'm hesitant to throw any code at the system to make that more convenient; i don't think people would use it.


The only thing i could think of in the now, is that we could create a special emoji for marking a post or thread as valuable. Maybe it's a gold nugget, or a gold script. This way everyone could vote on what belongs in the knowledgebase. But the problem remains that we don't have contributors. I don't know if making contributors lives even easier would help.

I personally planned to put a lot of elbow grease into knowledgebase contribution, but i have become progressively more busy to the point where i barely have free time to ride or work on ebikes.

whenever someone is about to navigate away from the thread they're reading, ask them to rate it. (technology already annoys us with this to ask for subcriptions to various websites when you try to close it, etc., so...)

Way too high friction. I personally avoid sites that do things like that.

 

[amberwolf]

No feedback has been provided on why we get no contribution.

Here's some, if it's helpful, but I don't think there's anything you can do about any of it, and I'm wierd and atypical and unrepresentative of anyone else, which is why I hadn't provided it, beyond this post.

So:

A long time back I wanted to create articles, instead of having to type the same thing over and over to answer questions, and had started making some starts (with thread lists for references) in the old wiki, but the wiki got hacked, I think, and never got restored, just removed, wasting all that effort of mine and others. I don't remember if any of the data was salvageable or not, so maybe it *couldn't* be restored. Doesn't really matter anymore, but it was very depressing and disappointing, and I still feel it.


Then a while back when the KB was in development, I was going to start working on KB articles, and had sort of done this with some of my replies to people's posts, projects, etc., so I could use those later in articles, but then you'd posted this, less than a couple weeks after it was even available,

Might remove the knowledgebase tab for everyone except editors/moderators because it's currently sad to look at.

and I didn't want to start making stuff that would just be hidden and not accessible to people because of a feeling, because it sounded like "if there isn't a huge kb then there's no point to letting anyone see what *is* there". Probably isn't what you meant but that's how it felt, and still feels, with things like the above "That's our last shot to get people to use it.", and one more I thought I'd read that I can't find right now, I figure I'd rather just suggest to people where they might find things or terms they can search for in the posts we have than to write stuff that will go away, so I didn't try to ask to become a KB editor.

If I knew *for sure* it wouldn't be wasted time, I'd be open to retrying the process, but I would be using the time I'd normally spend helping out in threads to do it, partly because the search, while better these days, is still not terribly useful for finding very specific info unless I happen to know a set of words I used in a post and can search for those. From what you've said about XF's limitations, there's nothing you can do about that, beyond what you've already done, which is a lot. It just takes a lot of time for manual re-reading thru threads to find things, or just retype it up from scratch and hope I remembered all the necessary details.


Additionally, the wolfybot project (or rather, trying to learn coding) is using most of my time that I could use for the KB articles (and any of my other non-life-critical projects), because I am still trying to wrap my brain around things (coding stuff) that I know could be done if I was a programmer rather than not even a hack. I can only do tiny bits of learning at a time because I run into frustrating walls all the time, with only AI to "explain" things to me, and mostly I get led in circles and lose what I was thinking so I can't continue until I figure it out again. :/ I don't think I'm much further along in learning than I was a couple of years ago, really.

Then I end up working on music to recover from the frustration of all that, because I can make the whole stressing-me-out-world go away while I'm making music (at least until I try to get people to listen to it, which is a whole other level of frustration, but that's not happening while I'm making it, so...I spend even more time making it as a way to avoid that, too).

Also much more worn out from work since this year started, as the new store boss that started then, and my direct boss, between them act just like my mom and dad did and i have to deal with both so it's lots of flashbacks to being a kid in that situation, which makes life just that much worse and harder to get past even once I get home....

 

[amberwolf]

whenever someone is about to navigate away from the thread they're reading, ask them to rate it. (technology already annoys us with this to ask for subcriptions to various websites when you try to close it, etc., so...)

Way too high friction. I personally avoid sites that do things like that.

So do I (if a site pops something like that up when I try to do something, I just add the entire site to the Ublock list and close it), but if people didn't actually respond to things like that I'd assume that sites wouldn't be built with it, and services wouldn't text and email you asking for ratings, etc.

Anyway, was just a thought--as previously noted, I'm full of those, good and bad.

 

[neptronix]

Responding to AW:

XF knowledgebase search: i also want this search feature, it's on the todo list for the KB, but there's no use building it without content.

Threats to remove the KB link for non-editors: I see how that might be discouraging, however, notice that i didn't do it. it kinda cast a bad look on the site to be displaying an empty knowledgebase though, so i don't like that.

Old Wiki: This software may have not been setup correctly in the first place ( not my work ), is extremely slow, may or may not have been hacked, and has really bad editing and organization functionality - worse than phpbb. A huge problem is that it has a manually cultivated index, therefore the index is never up to date, and misrepresents amount of content inside. Hosting it publicly causes the server to crash from overload. That's ultimately why it was removed. I still have a copy of it. If anyone was serious about moving the content over, i could put it online in a password protected directory.

Our KB software was developed as a response to the above problems with the wikimedia engine, and also finding out that other software wasn't any better. The only thing that risks it getting yanked is that it continues to not get contributions for a number of years going forward. Otherwise, it was designed to go for decades without any of the issues wikimedia or other similar softwares have.

 

[E-HP]

No feedback has been provided on why we get no contribution. One problem we know of is that the WYSIWYG editor is in some instances finnicky, but equivalent to the forum's post editor in ease of use, though it has more features.

1. I don’t know if I am accessing the actual KB. Does the tab navigate to the production KB?i see something about the testing area.

2. Hopefully the default sort of “newest first” isn’t used for the production KB. The KB seems to be set up like an encyclopedia, where the title of the articles are critical for finding the information being sought, so the default sort should be alphabetical. “Newest” is handy for an editor or reviewer I guess, but clunky for a user.

3. Not understanding the vision. Is the KB mainly geared toward experienced users? It seems like you need to have basic knowledge already to know what you’re looking for and how to find it since the categories are narrow. E.g. a newbie might look at an article on batteries and learn all about them and a bunch of details about cells, etc, and not read about BMS’s because they need to know to look at another article. If they simply googled ebike batteries, the resulting articles would likely layout the fundamental components to give the newbie a full pictures including touching upon things like the BMS.

4. Feels more like information than knowledge. Is that just the name coined by xenforo, or is it really supposed to record and convey knowledge? If it’s meant to store information, it has to be organized and easier to navigate than just doing a google search, so might need to revisit the structure. If it’s meant to convey knowledge, then it’s the wrong tool and should probably be scrapped.

 

[neptronix]

1. I don’t know if I am accessing the actual KB. Does the tab navigate to the production KB?i see something about the testing area.

This is the actual kb.


If you click on the testing area, it explains what it is. It's there for new editors to try things.
Other than 1 person who contributed a partial article, we don't have editors other than me. So i plan to keep this section in so people can try things confidently without messing up the existing KB.

2. Hopefully the default sort of “newest first” isn’t used for the production KB. The KB seems to be set up like an encyclopedia, where the title of the articles are critical for finding the information being sought, so the default sort should be alphabetical. “Newest” is handy for an editor or reviewer I guess, but clunky for a user.

That's the default sort. it's there to show you, in the right, what's new in that category. Otherwise you can use the search function or navigate through this on the left 'file explorer' like menu.

A paginated alphabetical listing of articles doesn't scale well with large knowledgebases, since it throws out the concept of categorizing things. In ES, there's a lot to cover, which requires either good search, or a heirarchical menu to navigate.

3. Not understanding the vision. Is the KB mainly geared toward experienced users? It seems like you need to have basic knowledge already to know what you’re looking for and how to find it since the categories are narrow.

The KB is geared towards everyone. The knowledgebase is mostly empty. We wrote the software and were hoping people would contribute the knowledge. The categories are narrow because those are just placeholders i put in to help people visualize what it would look like if content was in there.

E.g. a newbie might look at an article on batteries and learn all about them and a bunch of details about cells, etc, and not read about BMS’s because they need to know to look at another article. If they simply googled ebike batteries, the resulting articles would likely layout the fundamental components to give the newbie a full pictures including touching upon things like the BMS.

We don't have an article like that but there's no reason someone couldn't write one.

4. Feels more like information than knowledge. Is that just the name coined by xenforo, or is it really supposed to record and convey knowledge?

Call it what you want, we built this system from scratch and call it a knowledgebase.

If it’s meant to store information, it has to be organized and easier to navigate than just doing a google search, so might need to revisit the structure. If it’s meant to convey knowledge, then it’s the wrong tool and should probably be scrapped.

'use a google search' is how people navigate wikipedia and wikimedia based wikis because the search function is useless and there's usually no functioning index. The zeropress software is the opposite of that.

Here is an example of navigating a large instance of the software. I'll usually go through this left hand menu instead of search to find what i need because it's faster.

 

[E-HP]

This is the actual kb.
View attachment 379065

If you click on the testing area, it explains what it is. It's there for new editors to try things.
Other than 1 person who contributed a partial article, we don't have editors other than me. So i plan to keep this section in so people can try things confidently without messing up the existing KB.



That's the default sort. it's there to show you, in the right, what's new in that category. Otherwise you can use the search function or navigate through this on the left 'file explorer' like menu.

A paginated alphabetical listing of articles doesn't scale well with large knowledgebases, since it throws out the concept of categorizing things. In ES, there's a lot to cover, which requires either good search, or a heirarchical menu to navigate.



The KB is geared towards everyone. The knowledgebase is mostly empty. We wrote the software and were hoping people would contribute the knowledge. The categories are narrow because those are just placeholders i put in to help people visualize what it would look like if content was in there.



We don't have an article like that but there's no reason someone couldn't write one.



Call it what you want, we built this system from scratch and call it a knowledgebase.



'use a google search' is how people navigate wikipedia and wikimedia based wikis because the search function is useless and there's usually no functioning index. The zeropress software is the opposite of that.

Here is an example of navigating a large instance of the software. I'll usually go through this left hand menu instead of search to find what i need because it's faster.

View attachment 379066

Like building an ebike, maybe make a conscious decision that it won’t be a one and done project, and if folks don’t have time to devote to a lot of typing, change the approach so it still keeps moving forward.

The KB Hub Motors article mainly consists of links. It’s actually very useful, even without a bunch of explanation, as is. It’s an example where value is added without too much time invested in typing lengthy explanations.

If flagging good posts (reviewed by ES SMEs for accuracy) for searching isn’t possible, the KB could be used to capture that now instead, by category/stub. That might prove useful in the interim, and at some point there could be an effort to rewrite or even better, summarize the info in an overview at the top of the article, and still retain the links. The knowledge, or value added by the experienced KB Contributor team would be in that summary, and the gathering and vetting of reliable information within the links.

As an example, I probably wouldn’t spend time on writing up an article on BLDC motor testing, when TommyCat’s article in my mind, is the definitive reference. A link to the article is all that’s needed now, and a high level summary regarding how to use the article would be all I’d add as a summary.

Maybe Im the only one who bookmarks good articles when I find them for future reference. In my FAQ I’ve updated the links on motor fundamentals a couple of times, if I saw a better article, rather than attempting to rewrite the info in the FAQ. Links are easy to update.

If people added their best links to info for each stub, there will be immediate value. Those could be reviewed for accuracy by SMEs and the vetted links would provide an even greater confidence in that information. Then finally, a concise summary could be added to provide context regarding the stub category and information provided in the links. Minimal typing (which appears to be the biggest barrier), but still applying ES knowledge when discerning what information meets the quality standards that we’d like to maintain.

 

[neptronix]

I'd be happy with articles that just consist of cultivated links to posts. It's better than what we currently have.