Distributed ledgers of semi-structured organisational information

(Half* of this is likely indistinguisable from horse exhaust. I’m just not sure which half…)

In several talks, Vinay gave a brief overview of the evolution of databases from clean rooms to blockchains. But processing didn’t start with databases, there have been paper forms for a long time before. The parts that were worth the effort went into databases, anything that wasn’t ended up in a spreadsheet somewhere at best, or was completely ignored.

“computer says no” is what happens to human diversity in large organisations with remote databases.

While many organisations wish to believe they run off data, all to often in practice, they run off the stuff that the administrators have in their heads. How many organisations are run by PAs, with the bosses going to meetings, thinking they’re in charge, but doing what the PA says is next? Morlocks vs Eloi?

The “official” data may be in the database, run by the DBA somewhere, but decisions in a changing world are paralysed without the informal knowledge – if it wasn’t, it will be automated. Putting new structures into oracle takes time and is institutionally hard.

Everything else goes into spreadsheets, which get shared between people who need to see them, and hope nothing goes wrong. If bureaucracies run off spreadsheets, what happens when they get network effects?  The synchronisation of managed databases, but the ease of use of excel. Scaling up the efficiency of a well run office much larger, for information “everyone knows”.

Google docs help – you can share the same information, but if you’re looking to make decisions, you copy data from one spreadsheet to another and hope nothing structurally changed in the process they’ve been doing for ages, and hope there’s not been a change.

This stored information can be copied, and notified, in a way that is clear what happened and when. But the edit history on a google doc doesn’t really cut it.


What could happen?

There is nothing stopping any arbitrary spreadsheet signing particular values, and then putting the hashes, with or without the values, on a distributed ledger so everyone can know when it changed, and know whether or not they have the current value.

ONS published many such spreadsheets.

Here is the UK Inflation index as a spreadsheet, with an internal blockchain of figures as they update. Sheet 1 is the data with newest entries at the top; sheet 2 is the raw CSV loaded from the ONS published CSV (ONS publish new figures at the bottom, because statisticians), and Sheet 3 is the first sheet in a ledger form (there’s a macro “sheetchain” to run and update it – it could trigger on update, but currently that is turned off).

As that google doc is open, it’s available to any system that does https and javascript. Which includes the lisk blockchain, which lets you write your apps in javascript (and not necessarily in public).

As such, this is effectively ONS statistics published straight to a blockchain to allow smart contracts have up to the second information, all at the same time. Bloomberg for all, for free, for any data in any spreadsheet? Either publicly or privately?

If you want an auditable, logged, distributed access to information that’s locked away in spreadsheets, that is now possible. Which seems likely to have some level of effect. I’m not sure what they will be, but probably very little in the short term, but it might be fun to watch.


(tech note: It could be done on ethereum or any other platform, but lisk is javascript, which makes it easier to do. Javascript is annoying for technical purists, but it’s the language of the web. Ethereum may be python as to lisk is to javascript, but I’m not sure pure python is winning. But I was always JAPH.)

posted: 05 Jun 2016

How many teslas would it take to clean london’s air?

It’s almost an aside in this week’s Tesla piece about their “biohazard mode” air filters:

“it began to vacuum the air outside the car as well”


The Model X in the air safety test has an internal capacity of 77 cubic feet, and cleaned the air in 3 minutes.


How many tesla car filters would it take to clean the air of the London congestion charge zone, to the height of the Shard?

The answer comes out to be 880,000, which is 2 car sized filters for each person who works in the city of London (an area about an eighth the size of the congestion charge zone).

If we only care about air to the height of 10 storey buildings, it’s 86k filters, or one per 5 people who work in the square mile).

The spreadsheet is here if you want to make a copy and run the numbers for elsewhere.

It wont do everything (or anything that filter doesn’t do, other filters are available), but as a mass produced item, solar powered, that seems achievable, especially for areas with high particulate problems, and running full time may create other issues. They’re also probably the wrong shape for London buses (8000 in total, 1000 routemasters; 22000 black cabs).

The numbers are more feasible than I expected, when the question came to me at 1am. 

added later:  someone asked about cost. The normal filters start from under £2 new on Amazon. While size and quality will push that up by 10x or more, a sensible shape and volume may count in the other direction. I don’t know enough about care air filters to say…

posted: 15 May 2016