What do you mean? It's a thread in


IIRC I originally posted it to "Announcements (Altcoins)" but moderator moved it to tokens. Which I kind of disagree with -- while CHR token is now ERC20, the platform is going to be completely independent of Ethereum. But not worth arguing about, I guess.

Hi, I wonder if anyone here remembers me

https://chromia.com/

Chromia (by ChromaWay) is a new blockchain platform for decentralized applications, conceived in response to the shortcomings of existing platforms and designed to enable a new generation of dapps to scale beyond what is currently possible. Chromia is both a blockchain and a relational database. This means that decentralized applications (dapps) can be written in a way that is familiar to developers all over the world, whether they work on large enterprise applications, games, or smaller projects.

Whitepaper: https://chromia.com/indexwp.html

Programming dapps: http://rell.chromia.com/en/master/ https://try.chromia.dev/

Sia usage report #2.

Some time ago I tried out Sia UI, just uploaded/downloaded few mp3 files.

Now I wanted to see if I can use it for something practical: backup. So I used it to upload 3 files 28 GB total.

This time I used command-line interface as I run it on a server. I liked command-line more than GUI.

Upload speed: I don't have exact number but it took many hours to upload, perhaps 24 h total. I expected it to be faster as I'm using fairly powerful dedicated server with fast internet connection. But OK, it could be worse, recently I was a customer of an online backup provider which was MUCH slower than that.

Cost: I loaded wallet with ~60K SC, and right now it's empty. Thus at current price it cost me 60000 * 0.00000024 = 0.0144 BTC, or $6. As I understand, that's for ~1 month. In comparison Amazon S3 standard costs $0.03 per month-gb, thus 28 GB would have costed me $0.84 per month. So I'm quite disappointed, it looks like Sia is 6x more expensive than S3 standard, and 17 times more expensive than "Infrequent Access" option.

User interface: I'm quite disappointed with a lack of control over costs and redundancy, as well as a lack of status visibility. E.g. I'd like to know what degree of redundancy it has, what contracts are already paid for etc.

Stability: siad crashed twice during within 1-2 week time frame. Not cool. (I'm using this server for many years so it's unlikely to be a hardware issue.) I don't have any detailed crash report, some messages are in logs but nothing special around a time of crash. I'll try to get a better log.

Resource usage: I think siad memory usage went up to 3 GB during initial sync & upload, that's rather excessive. CPU usage is also quite high at times (e.g. during upload).

So, overall, I'm not going to actually use Sia for backups this time as it's more expensive than cloud storage providers. But it definitely looks promising.

It's very easy for an attacker to rewrite history. Suppose normal chain has blocks A1, A2, A3 which reference T1, T2, T3.

Attacker will create alternative blocks  A1', A2', A3' which also reference T1, T2, T3.

Perhaps he won't be able to create block A4 before T4 is known, but this has nothing to do with rewriting history.


When you analyze consensus algorithms you should consider behavior of a new node which joins the network and doesn't have any pre-assumptions. So, suppose a new node connects to one honest node and one attacker's node. Suppose honest node gives it a chain which ends with A1...A100 and attacker's chain ends with blocks A1'...A100'. How can a new node tell which of them is valid? Timestamps don't help here.

It can be mathematically proven that your block-base timestamps are no better than ordinary UNIX timestamps, assuming that node clocks aren't horribly out of sync.

Aha, so we just need to have a consensus to have consensus, got it.

You're begging the question. The whole point of this exercise is to make sure that transaction history cannot be rewritten. If you just assume that it cannot be rewritten, then you don't need all this Ti mumbo-jumbo.

Inclusion of Ti does not make it any harder to rewrite the history, as attacker has access to all Ti and thus he will be able to create a seemingly valid chain which references all Ti blocks but is completely different.

I think you've got things wrong, to improve security you need to reference alt-coin blocks from Bitcoin blocks, not the other way around. This is known as anchoring, it's already used in several projects (for example, Factom) and it can make alt-coin consensus as strong as Bitcoin. (See here: https://bitcointalk.org/index.php?topic=313347.0)

It turned out that it's so CPU-hungry only while siad is downloading the blockchain.CPU usage isn't a problem when it's fully synced. Perhaps siad sends lots of events during syncing or something like that...


So I deleted consensus.db and started over. It got stuck at 21664 again, but after some time went back to 205xx, then went further to 22040 and got stuck again for many hours (I tried restarting, reconnecting, etc).

I then tried a faster connection, and it helped: after 5 minutes or so it went further and eventually got fully synced.

To me it looks like there are several blockchain forks and siad is having a problem getting the latest one on a slowish connection.

So now I tried uploading and downloading some files. It worked, but I found several issues:

1. It's hard to figure out how much storing a file costs (or have costed, after the fact). (*)
2. I noticed it generates a bunch of transactions with wildly different sums, e.g. one is 0.02 S another is 5.4 S. Is that due to different prices or size is different too?
3. No information about redundancy.
4. Contract duration is specified in blocks, which is cryptic. I figured out that inter-block time is probably 10 minutes, then 6000 blocks is ~42 days.
5. The share box, after you opened it, it's impossible to close without restarting UI.
6. i find it strange that transaction list is not automatically updated, looks kinda confusing.
7. "Estimated competitive price" fluctuates wildly (I saw 5, 750 and some really huge number) and doesn't seem to be related to estimated cost on file tab. (E.g. estimated price is 750 S when cost is 220 S).
8. No notification after download is complete.
9. Most users will probably want to store files for a duration longer than a month, will it automatically refresh files? Can it refresh without downloading them? Do they need to be re-uploaded each time?

*: I tried to calculate it from balance changes, apparently it shows I spent 64 S so far to store 3 files 33 MB of files. This implies a cost of 1939 S / GB, which is quite a bit more than 220 S / GB estimate it shows now.

But actually there was a glitch: The first time I tried to upload these files I had to abort I had to abort it and restart siad at "processing" phase due to system thrashing on low memory, and it seems like money was spent even though upload wasn't completed... But it still doesn't explain difference in price, unless it was spending money repeatedly. Transaction list looks weird, there is a lot of small transactions, but also three spending 6.84 S (three transactions with identical values but different hash and time) and two spending 8.15 S.

Electrum, for example. It's quite popular and I haven't heard people complaining neither about usability nor about security.


Well, file + passphrase is a common scheme (user for ssh, pgp keys etc.), the idea is that if attacker steals a file it might be prohibitively expensive to crack the passphrase, or at least give owner time to react to breach of security.

In case with BIP-39, mnemonic is not encrypted, but instead seed is generated as PBKDF2(mnemonic, passphrase), thus seed depends both on file and user input.


OK, I'm just used to Bitcoin wallets which have access to public keys even when wallet is locked.


I'm using 0.4.8. My internet connection is slow, something like 100 KB/s, but it was enough for siad to download 2 gigabytes worth of something.

I waited for many hours. I saw it connecting to different peers, downloading something, but blockchain remained at 21664. I'll try to start over, let me know if you want to look at the stuck consensus.db.

BTW sometimes it shows 0.0.0.0 and 127.0.0.1 in the list of peers, what is that?

I recently tried Sia for the first time, I'll describe my experience with it as a user, maybe this feedback will have you to improve the product.

I found Sia via a reddit ad (I think it was a promoted link, although not completely sure), I got really excited because I'm actually looking for cloud storage for backup purposes, and other projects like maidsafe and storj don't seem to be ready, and I haven't heard about Sia before (even though I'm following news).

I am myself a Bitcoin user (and a developer) for 4+ years so I'm not easily scared with bad UX.

1. I downloaded Sia UI and launched it in a VM with 32-bit Ubuntu precise. Initial experience was quite good, it was easy to launch the app and UI looks great.
2. Block download is kinda slow, but i have fairly slow networking over here, so that was a part of the problem.
3. I bought some siacoins on an exchange and decided to transfer them to the wallet while it's downloading the blockhain.
4. So, create wallet experience... 29 word mnemonic, WTF? Most Bitcoin wallets have 12-14 words. 128 bits of entropy should be enough, maybe 160 bits if you want to err on safe side, but 512 bits are an overkill. You should read up on key size recommendations. (See below.)
5. Now to generate password I need to "unlock wallet", and it asks for password. What password? I thought there is a password separate from that huge-ass mnemonic as is the case with BIP-39, and the form field looked like one should type a small password into it, but having no other password (I first tried an empty one) I pasted this 29-word phrase and it worked. Well, it was unlocking it for like a minute, so it might be that KDF you're using is also an overkill, haven't seen anything like that in Bitcoin world.
6. I finally got an address and sent coins to it.
7. I noticed that UI is constantly using CPU even when nothing happens. `top` shows ~5 Sia processes besides siad, each eating 10-20% CPU, WTF? This might be because it's a VM without graphic acceleration, I dunno, but that's still kinda WTFish. Processes go away when I close UI and get back when I launch it again. CPU usage is really annoying as I'm using laptop and it makes fan noise when Sia is running.
8. siad got stuck with blockchain download at a certain point and I restarted it, this helped.
9. After siad restart wallet is locked again, so I needed to unlock it. It doesn't even show addresses before wallet is unlocked, why?
10. So I need to enter the large passphrase again. This is a major security/crypto fail. There is no way user will actually remember 20+ words, so he will have them in a file on his computer. So unless it's stored in password manager, it is as secure as just storing seed in Sia database. I recommend to look into how Bitcoin wallets work, they actually have pretty decent UX.
11. Also if you want Sia to be used outside of Bitcoin community, it needs to be even friendlier than Bitcoin wallets. Ideally no passphrase should be necessary just to try it out, it should appear only when user is serious.
12. siad got stuck at block 21664, no restarts help. Also it managed to download 2+ GB of data, while consensus.db is only ~100MB.

So this project looks very interesting, but I'm afraid it will take a while until it's usable by people other than devout fans.

A bit on key sizes: DJB recently posted a blog on key sizes: http://blog.cr.yp.to/20151120-batchattacks.html
Short summary is that 128-bit equivalent security should be enough for now, unless there is some weakness and/or attacker is able to attack many keys at once. 128-bit AES is not good, but if you're using a decent KDF 128-bit passphrase is OK. Otherwise, it probably doesn't make sense to use crypto stronger than 128 bits for keys if you use EC cryptography which is only 128-bit strong. Using 512-bit passphrase is kinda like this: https://s-media-cache-ak0.pinimg.com/originals/96/01/bb/9601bb26d756a9a18ae5cee170fa1188.jpg

Sentiment is a bulletin board which uses Bitcoin for a Sybil-attack-resistant message ranking. User which supports a particular message can sign its hash using one of his addresses which holds some amount of bitcoins and submit this signature to Sentiment. Sentiment collects all signatures, computers a total amount of bitcoins which support a particular message and uses that for ranking. The idea is that this ranking corresponds to the sentiment of Bitcoin users weighted by the amount of bitcoins they own.

http://sentiment.memeexchange.com/

Note: Signing a message with your private key might have some security and privacy implications, effectively the same as Bitcoin address reuse. Publishing signatures increases risk in case of use of flawed wallet software or discovery of ECDSA vulnerabilities.

Can you please stop doing this?

testnet3 is useful for testing software in conditions which are similar to those on Bitcoin main net. A crazy high block rate is nothing like this, so this makes testnet3 useless for this kind of testing.

It is interesting to see how software reacts to reorgs and things like that, not all the time, OK?

Same question. WTF.

Is back, seriously?

This demonstrates utter lack of respect to people who worked on other colored coins implementations like CoinPrism, CoinSpark, etc.

CoinPrism offered a mainnet web wallet since May 2014, and it already has quite a few users and assets.

But, apparently, it isn't for real until it is done by Meiri and his team.

ChromaWay is looking for testers and QA engineers  to test and improve our new web and mobile wallets.

Testers: part-time job, probably as little as 5 hours a week.

QA engineers: part-time or one-time consultation, depends.

PM me if interested.

It isn't. https://i.imgur.com/7uDvnWp.png

Looks kinda random.

Also, for freebitmoney.com statistics shows 181 impressions, while impressions column shows 921. Which is right?

Hmm, I don't see a way to see where you got traffic from.

There is Traffic Sources -> Exclude, but this list ins't sorted, so it's pretty much impossible to get useful information out of it.

Nevermind, it was a problem with my DNS server. Works now.

It's kinda strange it only affects a-ads.com.

Yes.

It says:


Your server sends this header in response:


I've tried it both in Chromium and Firefox, works in neither.

Yep, no relation.

BTW, http://chromaway.com/ is us. But aren't planning crowdsale or anything like that, just in case.

Looks like the site is broken now: it was unable to load stylesheets and stuff.

New web/mobile wallet preview: http://killerstorm.xen.prgmr.com/alex/cw/demo-eng.html
(running on testnet, doesn't allow you to issue coins)

Compared to old WebcoinX, it has much saner code base. It is broken into 4 layers:

1. coloredcoinjs-lib: a colored coin library, handles 'coloring', constructs transactions
2. cc-wallet-core: wallet core, manages addresses and does the bulk of work
3. cc-wallet-engine: a high-level wrapper on top of wallet core, hides all low-level details and can be plugged directly into UI
4. UI

I believe that this architecture will simplify creation of custom colored coins applications.

Say, if you want to use colored coins as authentication tokens, you don't need a full-blown wallet user interface, but you want wallet backend which is able to construct transactions, do SPV verification and so on. You can take cc-wallet-engine and create some custom simplified UI on top of it, without bothering yourself with implementation details.

This is still a work in progress, but you can take a look here: https://github.com/chromaway