There's more to blockchain than dodgy cryptocurrencies
There are CryptoKitties.
At the Open Source Leadership Summit in Sonoma, California, last week, The Register caught up with Brian Behlendorf, executive director of The Hyperledger Project, and had a chance to chat about the state of open source blockchain technology. Behlendorf took over the organization in May 2016 after the Linux Foundation …
Thems some ugly kitties.
Of course blockchains are going to be used for a lot of transaction-type data. That will make certain types of systems easier. However, if you think that there will be an explosion in blockchains that takes out the currently-existing structure of financial tech, think again. For one thing, it will take at the very least a system for keeping private keys that is private and understandable to the below-average citizen. That has been a problem that allows cryptocurrencies to be stolen, and the blockchain still relies on that structure. Until you have a good answer (and that answer must be short because nontechnical people aren't going to wait) to the questions "How can the blockchain system better solve the problem of fraud than my credit card company who I can call to deny payment?", "Who is responsible if my data, money, or whatever I put into the blockchain is lost or stolen", and "How can I use this system everywhere with virtually the same level of convenience, even if I'm disconnected from the network for some time or my wallet/blockchain-enabled-storage-device-thingy has broken down.", you won't see the systems switching over.
That's correct, but in addition to explaining blockchain, for it to take off it also needs to find a source of value. Certainly people can design a whole lot of things anew to incorporate block chain. But what is the quantum of any value? In the era of digital systems, does it avoid the need for data processing (no) does it avoid the need for storage and backup (no), does it avoid the need for reliable communications (no). It appears vulnerable to rather too many of the same forms of fraud as existing digitised currency.
I worry that I'm a Luddite in saying this, but AFAICS, blockchain is a sort of digital Esperanto. Cleverly conceived, logically designed, with enthusiastic adherents, but whose adoption is ultimately lacking sufficient tangible benefit to offset the not inconsiderable costs.
I wholeheartedly agree. My problems with the system, however, are not just limited to explanations. I don't think current blockchain solutions have good answers to my complaints, in that private key management, fraud, or availability/resilience are left up to the original authors of the code (bad) and individual users (worse). I respect those who will keep their private key nicely managed and stored; that's my approach and it leaves a lot of opportunity for fixing things and self-management, but if it's required then my parents won't know how, and they'll just leave it on the blockchain registry company system that they found with their google search. History shows that to not be a good idea.
For one thing, it will take at the very least a system for keeping private keys that is private and understandable to the below-average citizen.
No, the goal is not to make it understandable to some kind of "ordinary person," it's to integrate it in such a way that they don't need to think about it.
Cryptocurrencies right now are on par with PGP type encryption - about five steps too complicated for most people to be bothered. Unless either has a truly compelling case for its use, or is easy enough that it doesn't add to your effort, neither will ever really take off with the general public.
A good comparison is encryption using HTTPS. It's now pretty much ubiquitous, not because Joe Public decided that they really needed it, but because a large proportion of the industry as a whole decided to implement it, and eventually make it a default.
Blockchains right now are still largely a solution in search of a problem. Aside from edge cases where they are immediately needed I doubt that they'll really take off until the technology is rolled into other things so that it's more or less invisible. That won't happen until there's a really, really compelling reason to abandon what is in use now.
"Digital Esperanto" - very good.
"...he explained, noting that techniques for dealing with issues like network latency – which can mess up a distributed database – still have to be refined."
Is that a bit of hand wave re the CAP Theorem? Distributed databases aren't much of problem EXCEPT for the fact of a dodgy network.
Well, yes and no.
CAP, and other results showing fundamental limitations on distributed agreement such as FPL, give you hard upper limits. But even when you're working within those limits - eventual consistency with a high-probability bound, say - you may want algorithms that decrease average and near-worst-case times to consistency, provide better failure modes, and so forth.
So there's still work to be done even knowing that you can't have perfect distributed agreement.
If it works, it will slot seamlessly into existing systems, and the users need never even know it's there.
> If it works, it will slot seamlessly into existing systems, and the users need never even know it's there.
Exactly. Blockchain is just a key-value database with certain special properties. If banks and etc find it useful, it will take its place in their IT environments alongside other databases that have other specific properties. The concerns about liability and protection of certificates and so on, they apply to the banks' entire IT estates and presumably come under the same regulatory requirements and legal structures. Blockchain is just one particular tool among many that may be useful to certain organizations, if their cost-benefit analysis justifies it.
It's not even really a database, it's just an immutable* list of transactions.
* Except when it isn't.
"Right now, it's this Cambrian explosion kind of moment where it's worth mapping the landscape rather than being Teutonic about it."
There's your catch phrase folks.
How invigorating, as Teutonic being, to be related with trilobites and echinoderms.
Concerning that blockchain technology, I would personally wait until it reaches the Silurian period - I like my creatures with some backbone and that was notably absent in the Cambrian.
There's perhaps a number of problems that could be solved by adding a blockchain, but that's not saying they should be. Git, for instance, with replication is rather similar to a private blockchain and may often make more sense than mucking around with blockchain consultants who often have dodgy reputations for seeing everything as a nail for their decentralised hammer.
The only thing I can think of is an in-house blockchain to prevent the BOFH from getting carried away.
Biting the hand that feeds IT © 1998–2018