I think the following links and email addresses should be better known:

As you may be aware, Chas Lowes site has been sold and the shop will be closing at the end of this month. Mccarthy and Stone have purchased the site and will be giving an open presentation of their ‘proposes Plan’ on the 28th of March in Trinity Church between noon and 7:30pm. If you have views on how you would like to see the site developed, please do attend. The presentation is being made by The Remarkable Group who are handling the PR for McCarthy and Stone.

It seems that the actual business has been sold to Gibbs and Dandy, who are going to open a store in Gunnels Wood Park (SG1 3BH) ‘soon’. Watch this page for full details when they become available. It seems clear that land in Knebworth is worth a great deal more as a development for retirees than as a retail outlet.

I have been an elected politician for three years now. I have spent a lot of time in meetings, where people have said a lot of things. In some cases they have said the same things over and over but in many cases they have argued for the Council doing things. In nearly ever case, the focus has been on the benefits of taking the action. The implicit subtext is that this clearly exceeds the cost.

Most politicians would, if pressed, concede that all actions have a reaction, and, in reality there are very few changes that leave everyone better off. The hope is that the costs are low, and spread widely, and will be offset by one or other of the other actions taken which, cumulatively, leave everyone better off.

This is not unreasonable. we discuss many policies which cost very little in financial terms. A grant of £1,000 will cost the average person in the District less than a penny. It seems obvious that the average person would simply never notice such a cost. The gratitude and clear benefits to the recipient are, however, very clear to the grant committee. The argument hardly needs to be made that the award is worthwhile.

From an empathetic point of view, this is the right thing to do. But a quiet voice in my head keeps telling me that the resources that those 100,000 pennies represent have been diverted from what the producers of them would have chosen, to a charity that most of them have never heard of. Maybe if they had spent the time understanding the good work done by the organisers of this charity, they would have come to the same conclusion that we did. By examining the application, the constitution of the charity, the record of good work that it has done, we have saved countless hours of due diligence that could never have been justified if they were each considering dropping just a single penny in a collecting tin.

This is the argument made by Ronald Coase. That the practicalities of atomised decision making, are such that it makes eminent sense to delegate to a small group who can then arrange the joint purchase of services that in the absence of all the frictional forces operating in the real economy would have been impractical to agree.

The District Council collects household waste, and arranges for it to be disposed of. Every household may avail itself of this service. The cost is very modest. Private contractors would probably spring up if this service were not offered by the Council, and householders would, in most cases, sign up with one or other of the private contractors. Although I believe that capitalism is an incredibly efficient mechanism for supplying many goods and services, I acknowledge that it is very unlikely that households would save money.

There are certainly problems with local or central governments providing services. One of the biggest is the difficulty in innovating, especially when consumers no longer have the incentive to adapt their own behaviours to enable more cost effective delivery.

Coase’s Nobel Lecture.

About Steve

This is my latest corner of the web.

I used to have a blog at Tumblr. The old posts are still there. Tumblr has a lot going for it: spectacular free templates with all sorts of clever effects, the ability to auto tweet and auto post to Facebook. The problem is that there is no such thing as a free lunch. I don’t mind advertising when it is for things sane people might want to buy. The problem is that, increasingly, legitimate advertisers have come to realise that online advertising doesn’t generate increased sales, at least not enough to cover the cost of it. The end result is only people seeking to prey on the vulnerable find it profitable to advertise on social media. The problem is particularly acute, because advertisements are increasingly indistinguishable from my genuine content, resulting in readers assuming that I endorse products that horrify me.

The source code for this website can all be examined at github. I haven’t given a link to the actual source because if you cannot find it easily you’re unlikely to find it of much interest. Git, and sites like github and bitbucket are wonderful things. If you are remotely interested in geeky things, I strongly suggest that you take a look at them.

This can be edited in the Chrome Dev Editor and pushed back to github directly from the editor. That’s how I added this line.

Git submodules are still a bit of a mystery to me, but I yet again followed the instructions on this page and this time it kind of worked.

It finally dawned on me that git push and pull link a local work tree to a remote repository. You can’t run a web server on the remote machine, point it at the repository, and expect it to serve web pages, even though that’s what the repository (possibly) contains. Git gurus, stop rolling your eyes!

You can’t even checkout the repository on the local machine because then git push will complain that you’re pushing to a master branch that it active and checked out on the remote machine. I discovered that the hard way. My web server does run a git server process. I tried it.

Anyway, after a bit of faffing around (!) I decided to edit the hugo source web file source on my local machine (running msys2 under Windows), push to (because it is so awesome as a git server, and its free), then git pull the (built) website source from github to my web server directory on my dreamhost shared web server machine.

Now to add to my blog I create a markdown file, add it to the github pages repo, then git pull it to the dreamhost machine.

As a slight added complication, I run the DNS for this domain on another machine - a digital ocean dropbox, running mail-in-a-box. I discover I have to look in the dreamhost control panel DNS settings for the domain to see what IP address to use for the A record. I can’t just use the main (arago) IP address. This is not entirely straightforward. I had to setup a number of A records: the blog.stevehemingway record and the bare record (@), to different IP addresses. The DNS section of the domain management part the Dreamhost control panel gives the list of DNS records to set up. Having the mail server running on a completely different platform is not a problem: to keep this working, all that is needed is a good set of MX records, which Mail-in-a-box sets up automatically.

This is the sort of thing that makes setting up a website so much messier than it should be, and is the reason that firms like Squarespace will thrive, as they offer a complete, bundled solution for a fixed, modest monthly fee.

Who would have thought that something so simple could be so …. messy?

Update on 1 Mar

I am making headway. You can see the input and the output in the github repositories. My understanding of git has been improving. I now understand that a remote repository wants to be ‘bare’ - free of working files, so if you want the generated files to appear on a webserver, you should just push them to a remote repository, and run a git pull command on the web server. I think github pages must do that via some sort of chron job, but for me to refresh the site on my droplet, it’s easier just to ssh in and run git pull from the command line.