birmingham.io

Midland Information Cluster (mic)

This is one of the reason I want ‘the thing’ to exist. We duplicate effort because we don’t know the resources we already have, unless individuals happen to have been around when there was a conversation.

I can give you (or anyone) access if you want.

Do you need to know our Github IDs to do that? I could publish XML source and graphics for my UML diagram(s) there. It doesn’t feel like somewhere someone thinking of coming to work in Brum would look but it’s better than the nothing we currently have.

Happy to set up a simple markdown-based linear list (as suggested above – also happy if @rythie does it and I can just offer a few contributions). PRs will make it easy for people to contribute, and if/when there’s enough data in there, then we can slice and dice. (Or sling it into airtable.com).

I get that relational structure, etc can help, but a sensibly ordered flat file isn’t of zero value when we can still search it, so let’s get that moving, I think

Yes, I did work on a similar idea before but had to abandon it - it was hard to get the momentum to make it worthwhile :frowning:

1 Like

I saved a 2nd draft of my diagram as a .PNG graphic. I couldn’t attach it here so I put it on a page in my WordPress blog until I remember how to work git.

https://wordpress.com/page/andywootton.wordpress.com/8687

I know there are gaps so let’s see if you find them or things I didn’t think of first. Clearly there are things about living in Brum that have nothing to do with tech, so this information could be shared with anyone interested in being in Birmingham, even for a visit.

@dombarnes It was @timmy666 who mentioned his coffee blog Hi, I'm Tim Wilson from Code Club

I have 2 year old knowledge of a 2 hour talk on git and a couple of plays since, as andywootton. I’ve never even seen a Github Wiki but isn’t that what we want for storing data, as @rythie suggested? Can only organizations have Wikis? I imagine it’s Markdown-based, so e.g. a page for coffee-shops, if anyone is keen to get started and we aren’t duplicating something that already exists.

I think we are going to need lists of all the organisations (from Silicon Canal) and IDs for all the people in the community and all the tech we use and the connections between them. There are immediately privacy issues. I’d like people to own their own data and decide how it is shared.

We’ll need to agree a format for locations (I assume of a point, for now, to map) and times (for events) too. Perhaps this exists for the calendar. I’d favour using OpenStreetMap rather than Google Maps unless there are technical reasons not to, as I no longer believe Google have their customer’s best interests at heart and I don’t want to be dragged into a Chrome-hole.

Have I made clear what I’m talking about yet? Because I’ve been thinking about things related to this ‘problem’ for a couple of years, things that appear obvious to me may not to everyone else. Just as any book that is more than a sequential story needs a contents page that explains it’s hierarchical structure and an index to jump right to a page based on content, I think network structures like the interwebs require similar tools with more dimensions. Search can give us “GOTO” but the ‘semantic structure’ is also a network with one fewer dimension. Models with more than 2 dimensions tend to have to be viewed through multiple 2-D ‘projections’ onto the models, like we do with the models of architects, mechanical drawings and database schema.

What I’m trying to do is collectively build a speculative structure for a distributed Birmingham.io Information Cluster (le Bic?) that we don’t have yet. Some of the resources may exist, so it should help us to work out what we have. I volunteered to try to build the map of the virtual network and share it, as an agile computer science experiment and in the hope it will inspire others to provide the nodes containing the information that the model can point to. At some point the model should evolve into a live-updated index - I have no idea how.

Am I making sense yet? Any questions?

I’ve been convinced by Rich Hickey that all data structures are built of datoms which represent a single fact. In his Datomic database they are: “An atomic fact in a database, composed of entity/attribute/value/transaction/added. Pronounced like “datum”, but pluralizes as datoms.” A ‘transaction’ is effectively a time-stamp for when the current value of the fact is updated, and ‘added’ is whether the record represents an addition or a retraction’. Values aren’t replaced in position. The old value is ‘retracted’ and a new one ‘added’, so there is a historic record of previous values of every fact. It’s a bit-flip to delete a datom and allows history to be archived away later, out of the main data structure. I THINK it is similar to git. Obviously, in a mutable database we either log the deletion or throw it away.

So, ‘datoms’ could be captured in cells, lists, matrices, graphs or relational databases and disguised in free-format documents. I’m trying not to care where they are, just to point at them.

We can’t call it Bic. It’s a registered trade mark and we don’t have it in pink.

DROPs mic.

Can we have a BirminghamIO/mic repo please and can I do things in it?

Are there any team conventions in how to use git? Should I keep my UML diagrams in my own git and import the .PNG or work on a ‘fetch’ from the shared area? Does everything happen in master?

Re git on a public project - make branches and issue pull requests to master - never commit straight to it

Interesting. I know Lean people with a 5 minute build cycle who say “never branch”. They use conditional flags for experiments. They are a large team but not a public project so they can probably be more certain of quality than I can guarantee :slight_smile:

Who decides whether to pull, on a shared project?

Peer review, followed by a project maintainer merging the PR, commonly.

The problem here is that I only want to put the diagram up so that people can review it :slight_smile:

Post it here?

Also, rendered diagrams aren’t great for collaborative content - if it does go into some repo, including the source, too, would be great </patronise>

I will put the source in git when I remember how but who has KDE’s Umbrello? UML tools ‘in theory’ have an XML-based exchange format but I’ve yet to see a diagram exchanged successfully and I have no idea how we could merge parallel changes. I thought telling me changes were needed was better. Agile 101 - the conversation has more value than the artifact (but I’ve always had REALLY bad memory so I need to write things down/draw pictures.)

I put the PNG on my blog just so people could see what I’m talking about. I don’t think the diagram will be ‘the source’ in the long run, network discovery from reality could be, but ‘reality’ doesn’t exist yet, or where it does, we haven’t located it.

I’ve put this up here: https://github.com/BirminghamIO/bic

The first thing would be fork the repo and commit what you want to that fork. You can then share that fork here in the forum. It would be best if the PNG images were also shared in a easily editable source form, such as SVG with a method to regenerate the PNGs (i.e. a script).

If two people edit the same file, it’s typically the second person to push who has to work out how to merge the changes, which you can avoid by not getting behind master. If you need help getting started, message me on here.

1 Like

Is it too late to change to /mic @rythie? I realised that it is a registered trademark. I think it is also about the Midland area cluster, in Birmingham’s gravitational field, like the Metro Mayor.

Ok: https://github.com/BirminghamIO/mic

Great, thanks. We wouldn’t want to start a war with France; you know what those cheese-eating surrender-monkeys are like for taking offence :slight_smile:

I’m afraid UML “isn’t quite that simple”. A proper UML model isn’t just a set of drawings, it’s really a set of projections of an n-dimensional graph. That’s why you can’t use Visio, whatever Microsoft tell you. Anyone who talks about UML ‘drawings’ or ‘diagrams’ instead of ‘models’ is probably a charlatan.

Objects and connectors can be visible on multiple views at the same time. I could export as SVGs - I did because they re-size nicely; but Wordpress couldn’t load them so I gimped them to PNGs. Edits wouldn’t be reflected in the model because you’re just ‘patching the binaries’, so they’d be over-written the next time there was an export. Sparx Enterprise Architect, the UML tool I used at E.ON had ‘source-control’ of models in a database but I think it used locking.
KDE Umbrello is ‘basic’ and a teeny bit buggy but I’m following beggars can’t be choosers protocols.

Thanks for the advice btw @rythie. I could remember that I’ve started locally and pushed up and started at the top and pulled down and one was much better. I just couldn’t remember which. I think GitHub help tells you to do it one way and BitBucket the other. Apparently I’ve set up an organisation too, when I helped out at BCU. I wonder how I did that. I haz notes.

They’re in a ‘gist’ in my Github which is lucky because I’ll need to set up .ssh up on the box with Umbrello installed.

Update: I’ve noticed that you said “fork” not “pull” now. I’ve made some changes in andywootton/mic

Proudly sponsored by Bytemark