As I mentioned before, if I could get an AI to rewrite the forum into our existing Khooseller platform I would do it.
Well I opened a dev branch onto an AI session to do just this the other day.
it's crazy tech as I can keep 3 coding projects on the go at once and just about burn all my daily allowance doing that. C4P is just one of those projects.
Here is the current state of play - after a day or so with it.
We have the backend written and it is just starting on a new front end.
I am going for a more click to view replies at that level plus click to see those below a post or open another level deep as per Twitter/X way of viewing things. This promotes flatter replies rather than endless deep ones. You will still get to the deeper posts, but they need clicking down to, so most replies tend to sit at higher levels. Ajax loading should stop the endless scrolling etc. We shall see what it looks like and how it feels.
Does anyone use threaded view anyway? For me it looks like more thread is on the screen than post and there is a lot of redirection needed to click on the thread view and see the post. Not always very helpful IMHO anyway.
Flat view has the issue of working out who is replying to who anyway.
So hopefully the new look will sort this out.
Signup will use our existing user registration and validation
Instead of car, make model etc, we will use Tags organised into hierarchies. Basically the same effect but generic
Next thing is that I will introduce a basic editor so bold, headings etc can be entered. Rather than an HTML editor we will use Markdown - rather like Wikipedia or Github and convert to view. We will also allow image upload.
There is a question of moderation, but apparently there are some free AI tools by Google, Open AI that allow this sort of checking - nudity, profanity etc. which should keep the place clean.
I'm going to give it a break for a couple of weeks whilst on holiday, but this is just a heads up as to where we are going.
This will probably upset a load of people, but change is inevitable and the existing code is really really old and has to run on a server running older OS etc.
This will pull it up to speed and allow us to add any other stuff we like easily like News, etc. because that stuff is all part of our larger CMS. The forum is just another module within the system.
I am sure you will want to add your 2c
|
I'd comment if I could understand a word of that.
I'm sure it will be ok though.
|
I think it means things are gonna change.
:o}
|
>> I'd comment if I could understand a word of that.
Pretty much what he said.
|
|
I'm just happy that you are prepared to keep us going - thanks.
|
Thanks for your kind words.
AI is amazing though - since I wrote the post it has written most of the front end. Looks very different so not convinced myself, but considering it worked out the look and feel itself it isn't bad.
|
>> Thanks for your kind words.
>>
>> AI is amazing though....
>>
but can we be sure that it's really you making these posts, Stephen? ;-)
|
|
Claude here ... Stephen has been replaced
|
This is just an attempt to deflect blame if it goes U U.
(udders upwards)
|
|
Will it mean it'll be optimised for phone?
|
>> deflect blame
I’ve been rumbled. Zero ahead of the game
|
I don't understand 98% of the first post by car4play.
Does it mean the layout will be different?
Does it mean that the posts will appear in chronological order?
Does it mean that we will be able to edit our posts for an unlimited period?
If it does then I think it will be a good thing - otherwise we go back to my first three words.
|
I think one of Stephen's/Claude's points is that it will be easier to see who is replying to who.
I have been using Claude AI (from Anthropic) for academic research quite a lot recently as I get to grips with my PhD thesis. Some of his... sorry, its... results are pretty impressive.
Last edited by: James Loveless on Tue 10 Mar 26 at 10:05
|
One of my main concerns is that I have made a few changes that will not work well with the way the forum currently works, but I believe will be much better going forward.
- I am getting rid of subjects on posts. These are only on threads to encourage users to stay on topic. The new interface doesn't show them on posts anyway
- The new view will favour posts under the OP rather than long threaded ones.
So the first view you get on say one like Panic Buying will be the OP followed by the ones at the top level. Everything else is not initially visible. You have to click onto a post to see its replies - and in this particular case the threaded view here shows it is nearly 10 levels deep!
This new view discourages this kind of thing because it's a bit tedious to drill down, but I can see people complaining as to where all the posts have gone. I am hoping as new threads get started they stay more at the top level.
In fact some forums don't have a threaded view (mums net etc.) and all posts are just chronological after the OP which someone mentioned here they liked anyway.
I'm on hols tomorrow so not much to show until the end of the month. I'll put up an old import on a test domain and let you all have a play.
|
Have you resisted the urge to get it to write your thesis for you? ;-)
|
|
Another vote of thanks for all the work behind the scenes to keep the forum going. Have a good holiday.
|
"Have you resisted the urge to get it to write your thesis for you? ;-)"
To take this seriously for a moment:
University students face big problems with suspected plagiarism and AI has widened the scope for such suspicion. I have had training about the use of AI in research and it is clear that using text generated by AI is no-go, and may be dodgy anyway, in that AI can make mistakes.
What will always catch students out is the point in their (face-to-face) exam when they have to discuss and defend what they have written.
This is how I use Claude. I ask general questions on a topic, then follow that with specific questions about what has been written on a specific point, then look up the actual source and read it myself. In academic writing the sources of one's ideas always have to be attributed - i.e. shown in footnotes etc.
In my case, my PhD is so-called "practice-based", something which exists in arts rather than maths or science, in which the student writes/creates something (in my case a portfolio of musical compositions), then writes a thesis both drawing on the theoretical aspects of what I composed and linking that with received ideas that have already been published.
At present I don't think AI would be able to read scores and write a thesis like that.
The usefulness of it is that it saves hours of trawling through possible sources, either online or physically in a library, in the hope of finding a useful point. And occasionally it throws up something you haven't even considered.
Last edited by: James Loveless on Tue 10 Mar 26 at 12:59
|
My son has a sports based degree. He got a 1st. He now marks assignments of current students on the same course.
It's not a particularly academic degree so when a student uses words or terminology he knows they wouldn't know or use, it's obvious what they're up to. He can then speak to them and ask them to explain what they mean. They know they've been rumbled and have been served a warning.
Last year he was marking an essay which he thought looked familiar. It was his own essay which had been put up by the uni as a model answer for that assignment.
|
Forum rewrite - Robin O'Reliant
