Why is blockchain perfect for TW?

[Mat]

Having just jumped in on the cryptocraze (a token release for Sophia) and having read up a little on blockchains, I'm wondering if maybe this is the distributed network we need for sharing tiddlers and plugins and also enabling to monetize ones efforts if desired?

Here is an article that indicates it should be possible. It's just an arbitrary article, nothing special about it.

So, I'm just throwing out the question: Might blockchain be the solution for many problems in TW?

<:-)

[Dragon Cotterill]

Blockchain is a solution desperately looking to solve problems.

OK, the idea of the blockchain is sound. It ensure that changes based on previous data is valid. There are two major drawbacks, one in the blockchain itself, and the second applicable to TW.

Firstly the big problem is the need for continuous calculations. This eats huge amounts of electricity just to be able to validate the transactions. The more transactions you apply, the more processing you have to do. This is inherently not scalable when it comes to enormous amounts of transactions. Bitcoin is just starting to feel these effect now, and once people realise the huge costs involved, the I reckon it will go into free-fall.

Secondly, and more applicable to TW, you cannot change any previous data.

TW is essentially a data storage, and some of that data is likely to change. Yes that change can be factored in to the chain, but sometimes data entered is wrong or temporary and needs amending.

No, blockchain is not suitable for TW.

[@TiddlyTweeter]

Dragon Cotterill wrote:

Firstly the big problem is the need for continuous calculations. This eats huge amounts of electricity just to be able to validate the transactions.

Absolutely correct.

Amongst the biggest benefactors of blockchain for the money systems have been "server farms."

Both "Mining" and "Validation" have had fractional profit that in some ways dwarfs the actual transactions.

I don't think that easily translates yet to projects for human good.

My 2 cents

-J.

[Mat]

On Monday, December 18, 2017 at 3:08:31 PM UTC+1, Dragon Cotterill wrote:

Blockchain is a solution desperately looking to solve problems.

 

OK, the idea of the blockchain is sound. It ensure that changes based on previous data is valid. There are two major drawbacks, one in the blockchain itself, and the second applicable to TW.

Firstly the big problem is the need for continuous calculations. This eats huge amounts of electricity just to be able to validate the transactions. The more transactions you apply, the more processing you have to do. This is inherently not scalable when it comes to enormous amounts of transactions. ...

Note I have basically no knowledge on these crypto things, so FWIW:

I believe the energy consumption is relative to how deep the mining is and the size of the blockchain. As I understand it, ERC-20 tokens are not mined, but premade. And for something like tiddlers, the blockchaing should be very, very, very short, even for popular tiddlers. So would this not mean that energy consumption is negiligable?

Secondly, and more applicable to TW, you cannot change any previous data.

TW is essentially a data storage, and some of that data is likely to change. Yes that change can be factored in to the chain, but sometimes data entered is wrong or temporary and needs amending.

Well, my naive assumption was that the blockchaing could consist of the tiddler itself (or a hash thereof) and that you cannot change previous data (i.e the chain) but only add to it... which is done any time the tiddler is edited. The finesse here is that a tiddler could always be traced back, so if it says the author is Dragon we could see if it actually has been manipulated only by Dragon or someone else after Dragon. (and Dragon = Dragons public key). 

My point was that maybe blockchain could be a way to secure origin of a tiddler and trace its history - because this will be an issue when we eventually get a federated TW network going. Assuming we want traceability.

Total misunderstanding on my part?

<:-)

[PMario]

On Monday, December 18, 2017 at 3:08:31 PM UTC+1, Dragon Cotterill wrote:

No, blockchain is not suitable for TW.

Yes and No ...

The concepts and mechanisms of blockchains are definitely intersting, even in a TW context.

... I'm intentionally using the plurals here. There are many "blockchains" existing already and every "cloud-provider" has their own implementation. Not every single one of them is expensive to use. ...

I think, that dat-project [1] is a bit better aligned, to the TW usecase. ... BUT

and also enabling to monetize ones efforts if desired?

is an issue there.

There is a nice blog post, from the guy behind the beaker-browser : Cryptographically-secure change feeds in the Dat protocol, and on the Web  ...

Why the mechanisms are a thing, but the chain itself is not. ..

have fun!

mario

[1] https://datproject.org/

[2] https://beakerbrowser.com/2017/06/19/cryptographically-secure-change-feeds.html

[Mark S.]

It sounds more like you want PGP/GPG style digital signing, which would have less overhead than bitcoin.

After you've created a tiddler, you could click on an icon to sign it which would put a hash in a field, perhaps "signature".

In the context of something like Twederation, when TW's are imported the tiddlers could be verified against a list of known author's and their public key. Unsigned tiddlers could be rejected or quarantined so that they don't get passed on by whoever pulls from your TW collection. That would make the system more secure from spam.

-- Mark

[Jed Carty]

Using beaker browser makes sense for tiddlywiki, but I don't know what problem we would solve with a blockchain.

[@TiddlyTweeter]

I don't think Beaker Browser makes ANY sense for TW unless you want it to become even more difficult to use that it is already.

:-)

A programmers dream too far.

Just IMO.

[@TiddlyTweeter]

Right. Signing via PGP. Kinda makes more sense.

Though, BTW, I'm actually UNCLEAR why any signing is needed?

For WHAT?

[Jed Carty]

In that article there isn't anything that actually is helped by the blockchain part of what they are talking about with the possible exception of monetisation, but the monetisation schemes currently in place in blockchains seem to just devolve into rent-seeking, so I don't want to support that.

The distributed decentralised networking part would be simpler to implement using a distributed hash table.

[Jed Carty]

Saying that beaker browser is too complex in a conversation where adding a novel implementation of a blockchain is being discussed is a bit odd. Beaker browser is beta software that would let us create a distributed federated network, blockchains are a vague and poorly defined technology. More often than not the technology used to support the blockchain is what people mean when they talk about them.

The distributed censorship resistant social media already exists in the federated social network (Mastodon is the most famous way to access it, but it has been around for over a decade). A distributed method of connecting nodes together without a central authority exists in the form of distributed hash tables. Bittorrent has been using one for years.

The blockchain part is just a distributed database that is close to unfalsifiable. But it is still centralised! Everyone involved in a specific blockchain has the same database, that is the entire point.

So if your goal is to track the provenance of a tiddler than you can use a blockchain. But that doesn't prevent copying and plagiarism, it just makes it impossible to impersonate someone else or to claim that you did something to the chain that you didn't do.

We can make things unfalsifiable with normal cryptographic signing and not worry about the overhead of the blockchain. The only thing we will be missing is the history which I think is of very limited use and not all chains even save the entire history.

Sorry if this is a bit more than the conversation here requires, but I have been around far too many VC types whose eyes glaze over at the mention of blockchain without having any idea what they are talking about. They just see bitcoin speculation and think that means any blockchain implementation will make them rich.

[@TiddlyTweeter]

Ciao Jed

I think you so far into that world that you are so far from mine on this issue (not others, I think you great! I learned much from you.) you are in outer space on this.

Beaker browser is beta software that would let us create a distributed federated network

Absolutely. And go for it.

ALL I am thinking of is a normal web user. Coming to TW. I CANNOT BELIEVE that they are gonna adopt BB as any kind of solution to their issues. Its obscure. Complex. Weird.

Maybe if there could be a simple install that handled its virtues into a seamless app it would work out. But as an option it is ONLY for Tech Folk.

Best wishes
Josiah

[Mat]

I have to reside with Josiah that Beaker Browser would be a definite bottleneck if we expect people to adopt TW. At least if it requires people to install separate software etc.

<:-)

[ste...@gmail.com]

On Monday, December 18, 2017 at 4:52:35 PM UTC+1, @TiddlyTweeter wrote:

I don't think Beaker Browser makes ANY sense for TW unless you want it to become even more difficult to use that it is already.

What is so difficult about Beaker Browser? IMHO the easiest "use case" is the following: Create a TiddlyWiki, upload it to Beaker Browser, share the resulting dat:// URL with friends or family (who also have Beaker installed, and who can easily bookmark the URL). As the saving mechanism of TiddlyWiki supports Beaker, you can easily update your TiddlyWiki in the browser without changing the URL.

This way, you can e. g. share weekly photos of the kids or the cat with granny, without uploading them to Facebook or any other centralized provider. 

Cheers,

Stef

[Mat]

What is so difficult about Beaker Browser?

Well, could it be made to appear like a "normal app" that people install e.g on their phones? We do know they cannot use a BB based TW in their normal browsers.

<:-)

[Mark S.]

It's needed for Twederation to prevent injection of spoofed spam.

I will add that someone will object that in Twederation there is no danger because you are *pulling* all tiddlers. But in reality, in a shared environment involving more than say 10 users, you do not want to have to check and verify *all* the traffic that you may be pulling into your system. Having a system that approves tiddlers based on proven, trusted signatures would allow node traffic to continue even if you have a busy day and can't look at every incoming tiddler immediately.

-- Mark

[Jed Carty]

Do people actually use twederation? I haven't heard anything about people using it in a long time. Some of the other projects I am working on can be used for a node-based version of twederation that should eventually make it simple to have cryptographically signed tiddlers.

[Tobias Beer]

Hi Mat,

I have no actual clue, but for now all I see is a term hyped to promote bitcoin
among the IT savvy or those who hold such promoters in high esteem.

Not being convinced that bitcoin solves any of the problems one might think it's (t)here to solve,
you can imagine how that reflects back on the "blockchain" enthusiasm for me.

So, how is bitcoin not another monetary scam build on solid IT
with "mining" patterns rather similar to how actual banks
get to invent money out of thin air?

Distributed networks have existed for a while, so the real question is:

1. What's different about blockchain?
   
2. What TW problems does it help solve?
3. How?

best -tb

[@TiddlyTweeter]

Thanks Mark S. TWEDERATION needs signing so that you don't get crap or threats arriving? Right?

I am familiar with PGP. I'm not sure how many other TW users are. Whatever the system it needs be as automated as possible otherwise it could confuse a stupid person.

J, x

[@TiddlyTweeter]

Ciao Jed

I do not use it because to use it as I'd need to SEE the benefit.

Maybe its a catch 22.

TBH I always thought TWederation was your and Mat's EXPERIMENT, not a finished product.

Since I am not very savvy I was waiting till I saw something that demands I get into it because of its productivity already.

Best wishes
Josiah

[ste...@gmail.com]

On Monday, December 18, 2017 at 7:16:36 PM UTC+1, @TiddlyTweeter wrote:

I am familiar with PGP. I'm not sure how many other TW users are. Whatever the system it needs be as automated as possible otherwise it could confuse a stupid person.

I think it might be fairly simple, if the PGP keys are generated and stored within the wiki (and are used solely for this purpose). If you would try to reuse keys that were generated elsewhere, or even to involve keyservers, things might get a bit crazy, though...

~Stef

[Ste Wilson]

I returned to twederation the other day and put it on my testing stuff tiddly spot just because I'd asked about it recently and thought I'd give it a go. It was a fairly painless process. I like that others can comment, that I select what gets shared and who I get things off.
I don't like that when I share something it gets time stamped in the title and this breaks all links to it in my Wiki...

I think for uptake it needs to be a part of then core that is perhaps disabled by default but when setting up your asked to opt in.. This way a critical mass might be built quickly... At the very least a version of it need to be in the default plugin lib and perhaps promoted on the default getting started tiddler.

[Jed Carty]

Is that why people never seem to show much interest in using what I make? Like multi-user wikis and twederation and other things that have been requested, because I say 'experiment'? I am not sure why I find that so funny. I have just given up on answering questions about sharing tiddlers between wikis and simultaneous editing and the like.

[Mark S.]

You likely use encryption in your browser every day without even thinking about it. Google groups, for instance, is in lock mode.

GPG in TW would require the user to add a key to their TW's store whenever they wanted to allow the posts of some given author. It could probably be done in just a few clicks.

-- Mark

[Ste Wilson]

I'll be grabbing your multi user stuff next year... A new member of staff so I suppose I should share...
:D

[@TiddlyTweeter]

Ciao Mark S.

So long as its procedurally clear its doable and I'm sure tolerable. At lot of bottlenecks in TW arise because its not brilliant at contextual help systems that let you know what is going on.

- J

[Mark S.]

The problem with titles is that people aren't always very creative. Every title added to the system needs to be unique or eventually there will be clashes when your "My Tiddler" clashes with someone else's "My Tiddler". Possibly Pmario's Uni-link could be used to maintain links even when the title changes.

-- Mark

[Ste Wilson]

Ahhh. Back to the unique name title field /caption field goodness :)
Would it be possible to just create a date stamped field {{title}} in shared tiddlers?

[Jed Carty]

To return to the original post, as was just pointed out to me, git commit histories are blockchains from before blockchains were named that. So tiddlywiki was using blockchain-related technologies from the start.

[Mark S.]

I don't understand your question.

In TW, all tiddlers must have a unique title. Thus, in a federation, all tiddlers must have a unique (datestamped) title. It breaks a fundamental principle of database design, but there it is.

Makr

[Mat]

Datestamped title vs not

I think I understand Ste's question and I think it could be solved: The request is to have as simple titles as possible. When a date is baked into the title, it is not simple anymore. Also, if it was a non-shared tiddler first and you then publish it, it changes title or becomes a new tiddler, and your old links to the original doesn't work or are no longer to the latest version of it.

So, one idea would be to make ALL tiddlers have a datestamp in the title but to hide this part. Or to simply have half of the tiddlers ID be the title and the other half the datestamp, which is in another field. (IMO that's a good idea) In your local wiki, a link could be parsed so that IF there are two conflicting (simple) titles, then - and only then - does the datestamp come into question. The clickable link might for example have an intermediary step in the form of a small dropdown that lets you select which of the conflicting titles it is you want to open. My point is just that I think it is solvable - and I also think the request for simple titles is justified.

...or maybe *I* misunderstand something fundamental?

<:-)

[Ste Wilson]

You have eloquently and precisely laid out my question :)

[TonyM]

Jed,

I agree with your position here, even cynicism. What is unusual about block-chain is it is an algorithm who people talk about. Its entered the common parlance but almost 99% of people who use its name have very little understanding of most elements of the algorithm, thus a lot of talk about it is somewhat useless. Whilst more often they are talking about specific features that the block chain algorithm offers. Features which are the product of a range of other algorithms as well. As is so often the Case in computing we need to ask "what do you/they want achieved" and we will find the algorithms to achieve it, do not ask us to provide a specific solution/algorithm such as block chain, when they have not even clearly defined what they want from it.

For example Novel Netware's Distributed eDirectory is a fully replicated database build on transactions, and objects, which has being capable of many of block chains distributed nature for more than a decade. But I don't here this in common parlance.

Regards

Tony

[Mat]

Jed Carty wrote:

Is that why people never seem to show much interest in using what I make? Like multi-user wikis and twederation and other things that have been requested, because I say 'experiment'?

IMO... or I shoud say IMGuess, the first reason for why "so few users show interest in stuff what ANYONE makes" is that few users ever find out about new plugins etc. The number of people who read a post are just not that many. That is of course one of the very problems TWederation aims to solve. Another reason is that, yes, things are experimental in the beginning and that does scare people away. Very few want to take part in the development process (and surely that also goes for both you and me in most cases, no?)

 

I have just given up on answering questions about sharing tiddlers between wikis and simultaneous editing and the like.

Well, those who actually ask are clearly the ones who are interested. I basically always read your posts. You're one of the TWizards here.

<:-)

[Birthe C]

I also read your posts, I would certainly not like to miss out on anything.

 

Dificulty beeing to persuade anyone I know real life to use tiddlywiki.

Birthe