Today, theymos made a post on the ideas to improve merit system for a decentralized forum. It's amazing to see admin sticked with direction to build up a decentralized forum. From that post, one crazy idea popped up in my mind that I would start a topic, in which I present all ideas and plans of theymos with intention to improve the forum, but have not officially implemented.
It is obviously that we all can trace back the post history of theymos to get them, but it takes time, and I doubt that not all members known about them and really stored them for later use (such as stored them in Excell spreadsheet).
In my opinion, it will help us to easier to follow the potential of changes on the forum in next few months or years.
Without such topic like mine, sometimes we forgot what theymos posted and asked for opinion from bitcointalkers; or even theymos forgot what he planned to to in his To-Do List.
1) Merit system improvements for a really decentralized forumThis thought occurred to me recently:
If you wanted to implement Merit in a decentralized forum (ie. one in the vein of Freenet's Frost or FMS), you could do it in this way:
- Everyone can, from their own perspective, give unlimited merit to posts, and these merit transactions are put into files which each user publishes via the decentralized system. (Like a merit.txt.xz which every user publishes.) Unlike on bitcointalk.org, you can also give people merit without an associated post.
- For everyone who has merit, you download their merit-transactions-list, but scale down/up all of the numbers so that the total merit that they send is equal to the actual sMerit that they own. It might or might not be useful to do this via some sliding time frame scheme so that merit transaction amounts aren't just continually diminished over time as they increase in quantity.
- Apply the above step recursively, creating a web-of-trust-style merit network
Then every user has a subjective merit score for each post (sort of like the bitcointalk.org trust system, which was inspired by FMS). And if you wish, you can assign people to be merit sources from your perspective by sending them large amounts of merit directly; these might or might not appear in the merit-transactions-list which you publish.
2) Inline response of trust reportIt's a good idea, already on my to-do list, but it requires other changes as well. If the person who made the rating can delete it as they can today, then they can also delete the response, which isn't good. My current tentative thinking on how it should work is:
- Raters can retract ratings, but not totally delete them.
- Raters can add a short addendum to their ratings, and this can be freely edited, but the original rating can't be edited. (The point is to prevent conversations via rating edits. A rating should be one particular issue, and the response should be a response to this issue. Not 100% sure that this is necessary, though.)
- Additional limits on how often you can leave ratings may be needed.
3) Epochtalk (new forum software)The software is substantially complete. The main period of development was a while ago; the current work is mainly just maintenance & relatively minor improvements. Try running it yourself and you'll find that it's working, fast, and nearly feature-complete.
The things blocking a transition from the current software to the new software are:
- There hasn't been enough testing. I think that immediately after transition, a variety of small missed features, bugs, and performance issues would crop up. As a result, if the transition happened now (which is technically possible!), I'd expect the post-transition user experience to be poor for months while these things are fixed, which I don't want.
- I am the only bitcointalk.org sysadmin and on-demand programmer, and I'm used to the current software. Furthermore, I need to frequently make changes to the current software, but each change I make might require alterations to Epochtalk, which is problematic.
- The current PHP software, while ugly and sub-optimal in many ways, performs well, especially since I have extensively modified the backend to add features and improve performance. So I don't feel much urgency.
- The data-transition procedure still has a few known minor bugs.
We continue to work on these issues. I think that ultimately I may need to hire one or more full-time people, since a big problem is that the full transition is likely to create a ton of work which I won't be able to effectively handle alone.
The software is not vaporware (it's long existed in a runnable state, and is currently basically feature-complete), and is not abandoned (look at the git commit log). If anyone is unhappy with the progress, I invite them to take the Epochtalk code and create a competing forum with it; since they won't have to worry about the transition issues, they'd have a much easier time, and their testing will also end up helping us.
In short: If you want the software quicker, go run your own forum with it, and work to get any problems or missing features you find resolved via bug reports, etc. This would increase public interest, provide much-needed testing, and I might even hire you to work on bitcointalk.org when we're ready to do the final transition here.
4) Black screen when loading forum pages in chromeI can't reproduce this on any of my devices, so I can't effectively investigate it. But I suspect that it's a recently-introduced bug in Chrome's hardware-accelerated rendering. Disabling hardware acceleration might fix it; maybe chrome://flags/#disable-accelerated-2d-canvas ? (NeuroticFish reports this isn't it)
If anyone can figure out exactly what causes it or whether I can change anything on my end to fix it, let me know.
5) Bitcointalk.org collectible coinFor bitcointalk.org's 10th anniversary later this year, I was thinking that it might be cool to produce a limited run of 50-300 coins and give them out (maybe all-expenses-paid or maybe at cost) to users who meet some to-be-determined criteria. I'm not a collector, though, so I'm not knowledgeable about this stuff.
- Does this sound like a good idea?
- What features would make the coin especially interesting?
- What's the best way to get a coin like this made?
6) Badges for good reportersI think that I'm going to add some badges for people who report a bunch of posts. Maybe one for 300 good reports, one of 1000, and one for 5000. Does anyone want to make or find these images? They have to be exactly 16x16 px (like
), and if you find them online, then they must be copyright-free.
7) Ideas for topic ordering in altcoin announcementsThere's a strong incentive to spam replies to topics in altcoin announcements and a few other sections because there's a lot of activity and competition for eyeballs there. This spam should just be deleted, but I'd also like to remove the incentive by changing the default ordering of topics in those sections from "last post" to something new. (You'll still be able to click/bookmark something to get the old ordering.)
What are your ideas for new ordering methods? My current idea is:
1. For each topic, get a list of the distinct users who posted there in the last 24 hours.
2. Give the topic a score equal to the sum of each such user's total earned merit.
3. Sort by topic score descending, and secondarily by the usual last post time.
With that, you still have topics with recent conversation listed first, but it requires more resources to manipulate, at least, and newbies have no influence. Maybe it'd give certain users overwhelming influence, though.
I want to think about it quite a bit more, so I probably won't do it soon.
8> LN support for forum payments systemGrin is new and clunky to use. I don't expect many people to use it, honestly. But I needed to rewrite the forum payments system anyway in order to support LN in the future and to fix some longstanding issues with the old code, so adding grin support worked out nicely.
9) Ideas of uncensorable DNSIdea for uncensorable DNSI posted in that topic in 2010.
Still haven't given up on the idea.
The specific idea in that topic was basically Namecoin, but implemented sort of like a naïve colored coins thing on Bitcoin. Eventually it
became Namecoin. In any case, it has the censorability issues I mentioned in my OP.
10) Ideas of an anonymous file-sharing systemIdea for an anonymous file-sharing system11) Addblocker subscription for signaturesI've thought that it'd be interesting to publish an adblocker subscription for signatures. You can block individual users' signatures with eg. bitcointalk.org##.sig186520, and/or you can try to block signature ads generally with filters like bitcointalk.org##.signature table.
Userscripts are a good idea. Let me know if some small change on my end will help your userscript a lot.