-
Notifications
You must be signed in to change notification settings - Fork 5
Repository of Thoughts
but I think that this concept of collating a 'reduced' subset of patterns to help plan a new project is an important one.
So what I think you are talking about is adding new meta data to the collection. Each new collation, in essence, is creates a new tag to the original data implying a new use case for a subset of the data - thus each new collation 'adds to' not reduces the entire data set.
Yes and no, it would be great to rip out some pages from A Pattern Language and staple them together to create a new work that just contained the patterns that are relevant to the project I'm working-on, but a service that was primarily about bookmarking from a single 'perfect' set of patterns as a reference work belongs on a traditional wiki.
Such a resource would have many entry points, and would easily support other pattern languages and form languages without them interfering with each other.
This site should be the place where you start when you embark on any kind of project, at any scale, to search and connect with people who have done something similar, grab their work and then start something new. But also something that a community can use to share and develop their vision, to attract new collaborators, and to support a plan of action for change.
Can you in any way describe what you see here in your brain while you were writing this?
Ok, imagine there are more projects than there are people, this is certainly true for me, some projects are personal, are business, or need happen at a community level. The value of patterns is that they are the first thing you start with, they are like a specification that isn't specific, that you refer to along the way to ensure that you haven't lost sight of the purpose.
As an example, say there is something obviously wrong with our street, and as neighbours we mean to do something about it. Where we should start is by assembling the relevant patterns, this clarifies the needs in our own minds, but also gives us something to show to other people without getting specific about plans. The main point is that we should start with the patterns and work from there.
So if there is a website where I can search for "street" or "community garden" or "workplace" and find little collections of patterns that other people have already assembled for their own projects, I can just read them and learn, or create a copy to adapt into my own 'project language' - Is this a known term? Lev used it.
But this 'project language', no matter how good, is someone else's work, I don't just need to rename it then swap some patterns in and out, if I'm sharing this with other people, trying to enthuse them, then we need to be able to make the work our own, by editing the patterns to suit our own project, rewording them to add our own voice, or even translating into our own language - but in doing this we can't be precious or worry about messing up someone else's work, this isn't a wikipedia page. With a wiki page the process is a cycle of "perfect and improve", like an open source library, but this needs to be "fork, fork, fork", creating something new each time.
Yes there can be some route for genuine universally applicable improvements to get moved 'upstream', but this will be relatively rare and won't change work that has already been 'forked', the system needs a process where quality work is ranked and floats to the top anyway.
Can you describe the window you see in your brain? How things are placed? What you do first or last?
In my imagination it looks like this: each project is a 'story' in the system. So a clueless visitor arrives at a story, either from a search or a link, it appears like a designer's sample board, say a grid of photos with titles or short text, these are each a link to a pattern and you can follow them get more detail. The owners of the story can rearrange this grid and add some explanation for the project as a whole - This isn't just a nice organiser for patterns, it is a campaigning tool to build a vision for a community project - With the option to take someone else's vision and make it your own.
Initially the site would be primed with a single story consisting of free versions of the 253 patterns, and if you like, the patterns in this story can be maintained and improved wiki-style by a select clique of users that, among other things, trawl the rest of the site looking for stuff to bring 'upstream'. But this story doesn't have any special prominence, the stories that you find on the 'front page' or in searches are the ones that are ranked by the system.
Or, perhaps, name an app or web site that you think has some of these qualities?
Not without using buzzwords like 'curating'. I've never seen Pinterest because my employer blocks it, but maybe this is a model of the interaction.
There really is a need for a free version of A Pattern Language in the same sense that Linux is a necessary Free version of a unix operating system. Not just because keeping these patterns hidden in an expensive hardback book is preventing them being used to actually change the world, but because there is such an obvious need for third parties to be able to mix, rearrange and add to them for different situations.
So having a copyleft version of the original 253 patterns in a clean format that can be taken away and used for whatever purpose is a fine aim in itself. My experience tells me that this needs to be in some revision-control-friendly text format, like wikitext or markdown, and not html, but this is a secondary consideration. A related project is then building applications that allow people to work with the patterns, we can have some ideas for that, but first you have to free the existing work.
Some learning from recent experience: my brother is building a house (actually a boat) for his family, so I loaned them my copy of A Pattern Language, with instructions to ignore all the patterns where the scale is irrelevant to their project and to pay special attention to the patterns that at first seem unsuitable or a bit silly. Actually they didn't need any of this advice and saw the point of the book straight away, I even got it back afterwards.
Now, instead of photocopying pages, scribbling notes, making index cards or inserting post-its, wouldn't it be great for my brother, his wife and I to be able to collaborate: collate the patterns that are relevant, edit, add our own information and pictures, and even add new patterns - We could keep this private or make it available as a 'fork' for others to use. The point here is that although the patterns are structurally a great candidate for turning into a traditional reference wiki, how they are actually useful is as an edited, and particularly, as a reduced whole.
So envisage a place that does still have all the patterns as a big interlinked list of pages, but what you actually see is individual stories that consist of curated and reduced collections of patterns, "my garden shed", or "our street". As a visitor (since this is a website), you can browse these, or fork one to create your own, or merge patterns from different 'stories' together, edit them, add your own patterns, or start from scratch using the original 253 patterns.
These reduced pattern collections will each be comprehensible as a whole 'story', with each typically consisting of just a handful of patterns - The author can decide to display them all on one big page, or just have an index page with a few clickable images.
'Users' of the site will be able to explore and clone entire 'stories', each pattern will have an action like "add this pattern to my collections", or "ask to collaborate on this pattern". Don't worry about 'forking', or the site being full of outdated and 'less perfect' versions of patterns, there never will be a definitive set of patterns. Though individuals or groups can decide to curate a big list of 'sample' patterns specifically to be used by others to create smaller stories, but ultimately such a big reference collection is just another 'story' in the system.
Such a resource would have many entry points, and would easily support other pattern languages and form languages without them interfering with each other.
This site should be the place where you start when you embark on any kind of project, at any scale, to search and connect with people who have done something similar, grab their work and then start something new. But also something that a community can use to share and develop their vision, to attract new collaborators, and to support a plan of action for change.
Thanks for reading, I appreciate that most of this has already been said, but I think that this concept of collating a 'reduced' subset of patterns to help plan a new project is an important one.
the work still needs to refer back to the original corpus, both because it's a recognizable platform and communications asset, and because it would be awkward to work around such a big "hole"
...
But one thing will be critical -- an interface protocol, so that the patterns can talk to each other, and be used together in new languages. I see this as critical, otherwise we have two fatal problems -- one, the patterns can't effectively be re-used in any meaningful new combination, and two, the question of what a pattern is (a clear structural map with salient problem-solving information) will fall apart, and "patterns" will just be any aspirational idea people want to pretend something about. I have seen this happen way too much within architecture. (Magical thinking" - if it's a pleasant-sounding metaphor, it must be true. So must be phrenology, bloodletting - anything goes.)
The Git model. (as in Linus Torvald's new open0source protocol...) This is what Ward is doing with Smallest Federated Wiki.
In the older open-source model, e.g. Wikipedia, there is a central repository, managed by a proprietary entity, and everyone works on bits and pieces in open-source fashion. But this is curated by the committers.
In the Git model, the entire structure can be freely copied ("forked") and re-constituted by a separate team. Then they can work on it, make changes, and get the whole thing ("whole" being key) to operate in a different way.
Then some or all can be re-merged with the older whole, through a "pull request" -- or it can stay separate, and perhaps grow to be as large as the older whole in user community -- or larger.
This is what patterns have needed, as Chris himself has pointed out. A language needs to be applied to a specific context, and adapted to it. Only then, the patterns can be exchanged, and then re-adapted to another context. So you have "project pattern languages" that function as "forks" of other languages, then re-written. And maybe they are reunited with "pull requests" -- or maybe they remain relatively autonomous (but still shareable).
Note that natural languages work like this, not like the old model. There is not one central agency that controls languages, allowing individuals to propose additions, as in Wikipedia. (Well, except in France! ;)
Instead it's the Git model. Each of us can take the whole language, and work on it within sub-communities. We can add our own jargon words, or lingo, or nuanced meanings. Then those can re=join a larger community at some point.
I think this is an exciting model -- and Ward is doing very exciting things with it through Smallest Federated Wiki. What I think we can do is to pick up that thread, and as it were, "fork" his project!
My original drive - back a few months ago - was to put up a site of the APL material - clean & clear
Now - I'm less interested in that - for two reasons:
- The legal / personality issues. Takes some of the wind out of my sails to think that CA doesn't want it done.
- With the larger scope of potential here - of a current, distributed, mass-participatory pattern development space - I'm feeling that focusing on the APL work too strongly would detract from the overall.
... There's a whole bunch of material in peoples computers and files that they're ready willing and able to put up online.
So - I think that our focus should be to create a system that makes it easy for folk to put stuff online. Get the tech taken care of and let's see what kind of material gets put into it.
I think that we can use one of the open source CMSs as a start - let the first stages of work shake out any difficulties, and we can move to something more bespoke if need be.
Discussion on licensing: will go into its own page
How about a Creative Commons license on this work so we don't have to go around this again five years from now?
After the conversations of the past weeks, I'm realizing that there are a a good few project opportunities here. I'm currently categorizing them as:
- An online workspace around the original APL patterns
- A workspace for other pattern work in the realm of the built environment
- A workspace for context/conflict/resolution oriented work in a broad range of areas (Each of these have a number of variations)
From our conversations, I'm hearing that there's interest in all of the above - some of us lean more toward 2 or toward 3, but all want to move beyond just 1.
Even so - there's a good amount of energy around getting a good (yes, legal) treatment of 1 online. I see it as a good stepping stone toward both 2 and 3, and a good targeted project to begin with.
My intent is to create a workspace that will use a fair-use-compliant skeleton of the original APL patterns to provide structure for a workspace that will allow users to comment, upload photos for and against, and link to projects and examples that use the patterns. In moving toward 2 or 3, I'll be willing to upload any other body of pattern work (under any reasonable license), and will provide tools for users to upload patterns as well.
Technological conversations are happening on the development list (open to the public - subscribe at http://groups.google.com/group/pattern-repository-dev). Theo has made an online version of the original APL work available for internal purposes - that's going to be very helpful. I'm investigating using some of the same technologies that Ward is using on the Federated-Wiki project, primarily so that in the future I'll be more able to understand and contribute to what is happening there.
I'm hatching an idea for a simple website to solicit input from the wider community.
Idea is to show each pattern with a short description (thinking this could even be the short description from the original) and original photo, and allow folks to upload/link images that support (or weigh against) the pattern, as well as add their own commentary.
Seems to me that is pretty close to the bare-bones needs to start generating new content around the original patterns. Would be relatively easy to allow creation of new ones as well.
Desirable content would be:
- number
- name
- lead image
- ...other images
- intro text
We heard Murray Silverstein describe the process of creating the original book. He described taking over a living room in south Berkeley and putting hanging folders from one end of the room to the other, arranged by scale. The names on each folder could be changed easily, and many of them started off unnamed. As material was collected - images, articles, studies - they'd get dropped into the appropriate folder. Sometimes folders would get merged. These folders ended up being the raw material for the book.
In that time, a process like this could only happen with a group that shared a location. Now, a very similar process can take place in the common location of the Internet. I think that is what we are creating here - a system of virtual hanging folders, the ability to collect material - images, links, blocks of text, 3d models, movie - and the ability to select, cull, refine that material into something of a 'current canonical' form that is the 'best of' the raw material.
The best form to present the full folder and to present the 'current canonical' will be driven by the medium of viewing - a screen, a pad, a phone...