This is a page about building a software system to support CategoryDecisionMaking, specifically GroupDecisionMaking, with application to a particular complex, many-person, many-year dialogue.
While discussing WhatCommunicationSoftwareToUse, we came to the conclusion that no one off-the-shelf piece of software does everything that we want this system to do. We hope that off-the-shelf software (Compendium or some WikiEngine?) that covers many of the requirements will be adequate in the short term. We hope that the remaining requirements can be covered with a reasonable amount of programming.
I have been gradually fleshing out a project on developing a long-term strategic vision for a group that I sometimes frame as the U.S. sustainability movement and sometimes as the the American Left. For convenience, I have given this project a working name: the Convergence Institute project. The project has three major components:
1. A residential center for Fellows of the Institute, scholar/activists from diverse movements and progressive traditions.
2. An ongoing set of dialogues between leaders and thinkers from various progressive movements organized and hosted by the Fellows.
3. An electronic community and decision support system to help integrate these various dialogues over time and space.
I am writing this group for help in thinking about the last of these three components.
Assume that we have a Convergence Community consisting initially of a dozen or so and ultimately of a few hundred or thousand people. These people are trying to figure out how the Left, broadly defined, can put the world on a path to a just, prosperous, and sustainable future. This is treated as both a political/policy question and a technical/economic question. Assume that there is a reasonable recruitment mechanism for people likely to have something intelligent to say, but that these are very busy eople, not techies, who will not keep coming back unless the community provides value to them. The conversation takes place over a decade or more, with people joining and leaving and many new ideas and approaches being discussed, tried, shot down, and compared to real-world experience and historical data.
In order to keep track of this conversation and let people join in late, rely on a piece or system of software. My current and very tentative fantasy for this software is something like this:
In terms of visual analogies:
The dialog map and the related modeling system or systems would need a reasonable versioning system (modeled on the Current Version System?) that enables people to see how the dialogue or model has evolved over time, and to save and name model versions, linked somehow to the current dialog map.
In terms of a web interface, ideally people who have a password would be able, through their web browser, to do anything that they could do with that program if they were local except delete it: see the output screen, press buttons, type into dialogue boxes, etc. They should be able to mess about with files to prepare data for input to the simulation or output for presentation. Finally they should able to download files to the folder and upload files from it. Archived models and dialog maps should be accessible but not deletable.
I am a lawyer and do not actually know anything to speak of about IT.
1. How much of this is this possible with a sane investment of time and energy?
2. How much of it can be done with currently available software, ideally open source, but commercial in a pinch?
3. Assuming that I will not be able to start with a system that does all of this, or even most of it, what would make the most sense as an initial kernel from which to grow the system over time?
4. Might there be interest in people working to grow some existing open-source system, such as Compendium, in the direction of this larger vision? How might I find out?
5. Are there other vital features of a community support system that are missing here? Identity and resource pages for participants, maybe? What kind of tools might help to link the dialog support system more immediately to the daily needs and concerns of the activists, scholars and opinion leaders who compose the community?
I started out by assuming that this vision was absurdly over-ambitious not only in its sum, but in most of its parts. However, I have become increasingly convinced that almost all of the pieces I describe have out-of-the-box standard solutions if I can only describe them correctly, and that the real challenge will be one of integration and establishing a reasonable growth path.
For example, i was talking to my housemate Martin a few days ago about being confused about whether everybody who logs in would be working on some central model, or if they each had their own conflicting version of the model. Each approach seemed to lead to intractable problems. And he said “Oh, like CVS.” “Huh?” “You know, the Current Version System people use to design open-source software.” Well, of course i had never heard of the Current Version System, but it was immediately evident that a whole lot of people had put a lot of time and energy and thought into what was essentially the problem i was describing, and probably come up with a pretty good solution.
There is a family of programs intended to support mind, argument, or dialog mapping. In fact, there are books about it. Keywords that if handed to your favorite search engine will help you find these include: Mind mapping (MindMap), dialog mapping, argument mapping (MappingArguments), argument visualization (ArgumentMap), collaborative tools.
Compendium is a pretty cool product for dialogue mapping you can download from http://CompendiumInstitute.org/ (not yet quite Open Source).
Rationale “argument visualization software” focuses on mapping arguments rather than dialogues. http://austhink.org/
“Decision Explorer” is a pretty expensive product that does dialogue mapping but also integrates some decision-making tools. http://banxia.com/demain.html
Dialogue-mapping or mind-mapping software shades over on one hand into into decision-support software that offers decision trees, critical path analysis, and so forth; and on the other hand, into groupware like FacilitatePro and GroupSystems’ Think Tank to support distributed meeting process rather than, or in addition to, creating a shared mental map. (The latter are quite expencive and I am not aware of any open source equivalents.
There are a gazillian programs that support decision trees and project planning, but I know very little about them yet, and don’t know which might be open source.
There are at least six different programs that look a lot like the “system dynamics” part of what I have just described, and two of them, Stella and VenSim?, have been used in curricula aimed at teaching the system dynamics way of thinking to high school or grade school students. They were built by people who wanted to do just what I wanted: to enable ordinary people to build and understand models of complex systems. VenSim?, http://www.vensim.com/, is free in its basic configurations.
The best-known global models, World3, (which supported the “Limits to Growth” books) is available in both Stella and VenSim?, and many other models as well. So my current notion is that, instead of being just one program, we have 2, but the programs support each other in a similar way as under the unified vision. Of course, World3 does not include a political model, and it has very little sub-global detail, but it does suggest some of what might be possible along these lines. The idea is gradually build the model up parallel with the dialogue map, allowing people to run their own simulations, tweak the parameters, and see what happens – and to tinker with the guts of the model as well.
For the first few years I see this as a small community of invited friends. Somewhere down the line I would like to look at the much more challenging problem of letting anybody do it. But by then there should be money for it that does not come out of my own pocket.
In terms of a web interface, . We have programs like:
Compendium can save its entire output as HTML, so I think it may be possible to do a “read only” version after each modification that could be accessed by anyone with a password, like a regular web page. I am not quite sure that this includes all the database-linking facilities of Compendium, though I see no reason in principle why it couldn’t.
In terms of starting with off-the-shelf software and growing in this direction, my best current sense is that you could do a lot of this with Compendium, VenSim?, GoToMyPC?, and some kind of group software that supports threaded dialogue and easy file and content manipulation – if there was something open-source to provide roughly the functionality of a Yahoo group, that might do the job. Or should it be more wiki-like?
You’re a very important person, your work is crucial, and I do not want to see your time wasted.
Properly answering your question will take a while. We tend to write here sporadically, and according to whim.
Your question is the sort of question that we like to answer. We do have insights into what you are asking. There are even scattered answers to several sub-questions throughout this wiki.
But there is no guarantee we will write the response.
You have provided your e-mail address.
In the event that we respond properly to this question, I will make sure that somebody emails you to let you know that we’ve responded.
I am trying to think of where best to take this question…
…are you okay with calling people on the phone? I have a few people you may want to call directly, to get some answers. No fees. (We understand the gravity of your work.) I am available for free consultation as well. My phone number is 206.427.2545.
For CommunityWiki: “Where can we best send this question?” Or, if you have the time to start a response, …
Yes, I would be very happy to talk to folks on the phone, to call them or have them call me. My cell is (510) 507-4820. Folks are welcome to call any time between 9AM and 12 midnight, Pacific time.
And I would be happy to talk to you – what is a good time to try and reach you? You can respond here, or by email or phone, as you wish.
Welcome Andrew! I will take some time to think about what you’ve written above and try to give some useful feedback.
Thanks! Looking forward to it!
Welcome to the CommunityWiki. Your vision of the software you want brings together many of the things we’ve been talking about here. Such as MappingArguments, DebateTool, WikiDebateBase, WikiDrama, FracturedDemocracyBuilding.
Perhaps it would be good to put your vision of software on its own page. Would you call it DecisionSupportSoftware? or DialogMapSoftware? or something else ? I’d be happy to make that page for you.
These sorts of questions – “I want to build a web site that does X and Y and Z. Doesn’t there already exist software to do that? What is it called, where do I get it?” – are exactly the sorts of questions that the SoftwareBazaar was designed to handle. Alas, I haven’t attracted enough people yet to make it a really useful resource yet – perhaps you can help with that problem.
Right now, InkScape is the best open-source program for drawing diagrams and maps. LionKimbro and I have discussed building a kind of wiki that makes it easier for people to collaboratively edit such maps and diagrams. (linear diagrams can already be collaboratively built using the EasyTimeline).
Thanks, David!
I would like the map to work like the table of contents, or index – to be a natural way in for people entering the dialogue at a late point; a place to hang a wide variety of stuff, from budget information to historic timelines; and a want the dialogue map ultimately to evolve naturally into an actual model of the process under discussion.
But what I want right now is something that exists, that I can get up and running in a hurry, that can be used to post a map of the first discussion we host on the web within the next three months. I am leaning toward using the open-source dialogue mapping software Compendium, because it has a bunch of these capacities already built-in. But I admit that I am very worried about investing a lot of time and energy into a product that may not ultimately be the right one. Specifically, as of this moment Compendium is feature-rich, and well-designed to be collectively used. But it is not well-designed to be collectively maintained. People are working to change that, but i have no sense of the likely timeline, and it is not an effort to which i can contribute directly.
What i think i will do is install Compendium on my machine, start playing with it, and write a review here somewhere. Then I will look at the other software people recommend, write up pro and cons of various approaches, and invite comment. Please do continue to send software ideas my way.
One thing that I can say offhand is that the description Andrew gives above of the system that he envisions reminds me of some of the concepts in Wiki:VisualTour. Wiki:VisualTour was an idea that was implented in the world’s first wiki, that allowed the wiki to drive a visualization pages and connections between them.
So, when I read Andrew’s concept laid out above, it seems, just off the top of my head, that the “core” of the system could be a wiki engine, that drives a network visualization of pages. Then the tools for things like decision trees, project management tools, could be added as some type of module, or built in to the wiki engine itself. I think it can work with wiki. If you built on top of a wiki engine, I don’t think you’d have to bother with anything like GoToMyPC?, or other remote desktop tools. The old Wiki:VisualTour idea makes me think that an OpenSource wiki engine could be modified to provide the framework for what Andrew is describing.
In ArgumentMap, LionKimbro linked to http://www.standpedia.com/, which appears to be a wiki-driven structured argument mapping system created by a private company.
In my humble little opinion, the optimum, based on your stated intended usage would be to build this based on existing OpenSource software technologies, and attract a community of people who can help develop it, maybe also sources that are interested in funding the development (sources that are enthusiastic about assisting the development of OpenSource software tools). But, this also has the potential for being more time consuming.
I don’t exactly what type of feedback would be useful for you, but perhaps CommunityWiki could help you by isolating your wish list/description into it’s own DocumentMode page, and then creating sections that show what we know already exists that fits or comes close to fitting your descriptions, both from pages within CommunityWiki and elsewhere. And then on the same page, or a connected page, we can also brainstorm and execute some TheoryBuilding around your ideas based on the wide angle view we can collectively take of what is out there. These pages could then also possibly spin off into their own wiki for your project, whatever direction it goes in. It could turn out that some of the ideas we have on pages like ProjectSpace, and maybe some concepts in CollectiveProblemSolving could work for your frankly awesome idea.
Sam, thank you so much for these thoughtful comments. I agree that Visualtour and Standpedia sound like good candidates to asses next to Compendium as candidates for the initial nucleus. I am not quite sure that I fully understand what a DocumentMode page is yet, but to the extent that I do, it sure sounds helpful. What input from me beyond what I have posted above would be most useful in getting this next step started?
Dear Sam-- I took a look at Wiki:VisualTour and the documentation made the following claim:
“The VisualTour? was a relatively easy way to look a few clicks ahead in your wiki browsing. At the bottom of every page is a link to the tour which picks up right where you were browsing (“or take a VisualTour?” in the bottom). The tour was inspired by suggestions to VisualizeTheWiki? (There you will also find scripts which visualize wikis).”
This appears to claim that clicking on the VisualTour? link on the wiki takes you back, not merely to the entire map, but to the specific portion of the map that connects to that wiki page. I observe this because this is something that i want to do as well, and because Lion suggested to me that while the technology to link from graphical objects to web pages is now a no-brainer, the technology to link from pages to objects embedded in a big page is not there yet.
I have a general question that i would like to address to people about what kind of help i need to make this fly. I know i need some basic assistance in setting up a web server and learning how to link the web-based software and any non-web-based software. I think that has to be someone who is physically present. But when we start thinking about actual development of new open source capacities, is this still true? Or would it make sense for me to show up in all the forums like this one i can find, do my little song and dance, and ask if there is anyone who would like to be involved? In your experience, how important is physical interaction and face-to-face contact for (a) clear communication and (b) maintenance of commitment to and enjoyment of what will certainly be an all volunteer effort for some time to come?
One of the reasons that i am leaning toward Compendium at this moment is that there is an existing community of people who are doing open-source development on it, and i have already established that a number of the capacities that i regard as missing are on their development wish list, and i know that I am not going to be adding any capacities to the software. But if i found one or more people with the technical skill that were interested in working on the project, i would certainly be inclined to substantially defer to that person about a development path, given a shared understanding of the ultimate goal and near-term constraints.
On the other hand, I am someone who, given dialogue diagramming capacity, can intelligently diagram a dialogue, and given a modeling capacity, can intelligently build models. So i like to think that i can assure someone who puts time and energy into building such a capacity that it will be put to good use, in at least three senses: use that is technically proficient and credible; use that is carefully tailored to provide politically significant input to processes with national and global strategic significance; and use that will, if it is effective, serve the common good.
Please consider everything that I am writing in here as just brainstorming with you. I don’t want to send you off on wild goose chases, of course. And in fact, based on everything you are saying here, it looks like your best bet, in my opinion at this time would be to stick with compendium, and maybe use the version that can be shared via Jabber, that is referenced here:
http://kmi.open.ac.uk/projects/compendium/download/collab-compendium.html
I say that because it looks like compendium is coming the closest to what you want, and if they are already working towards the features that you want, and they already have an open source way to link Compendium-to-compendium via Jabber, then this makes progress on what you need. Compendium can also share databases asynchronously, which is neat, though is not a real-time sharing.
If compendium seems like the best candidate to you, then joining their developer mailing list, and building relationships with as many of the people involved seems like a good direction to consider (meeting up with them F2F, inviting them to conferences, etc).
I hope this helps. Again, it is just some thinking out loud and brainstorming by me. Just my opinions, based on what you’ve written here. Hopefully some more Community wiki people will chime in, too.
Before I create a page or pages as I suggested in my last response, I am going to follow LionKimbro’s lead and see how he responds, and what directions he suggests, or at least give him some time to respond to so that I don’t end up unintentionally hijacking it. Because he is frankly more knowledgeable in this overall area than I am. Plus, I’d like to give anyone else a chance to give input too, as there are a lot of people with knowledge in this area active here.
Down at the bottom, if you click on Administration, there is then a link to the Index of all pages
This is my best current thinking: 1. It does make sense to start with Compendium. This conclusion is based mainly on four factors.
2. I will need a technical support person who is willing to work with me for at least the first year or two that the project is underway. There are too many decisions here of which I simply do not understand the full implications. The Jabber/VNS decision is a solid case in point. I should probably have this person before hosting the first dialogue, which I would like to do no later than February.
3. The technical support person does not need to be local, but probably does need to have an interest in the underlying dialogue and the challenges involved in creating a useful map and building a lively community connected to that map. That person could attend the in-person dialogues via Skype videophone or some similar technology. If they chose, they could take on a visual note-taking role, though this is not a requirement. It would be a big plus, but not a sine quo non, for that person to also be interested in joining the Compendium development community.
Is there anyone here who:
AutoWiki? is An imaginary but (i think) not terribly difficult feature to add to a piece of dialogue or argument mapping software such as Compendium.
I have now identified several different fora in which there are people who want, or are talking about, or say that they intend to build, or claim to have built ArgumentMap or dialogue mapping software that can be edited collectively (asynchronously) over the web. At least two of these, Compendium and Standpedia, also currently produce unique web tags for each node of the argument. So let us suppose (though I think this is not quite true yet) that we start with an application that produces a map that is editable by any group member (with good “restore” and history functions!!). When the AutoWiki? feature is turned on, we require that, as the map is constructed, each node and line be named meaningfully, using the common wiki CamelCase naming convention. You would need a special proceedure to add such names if you turn it on for an existing map without them. The meaningful names would be used in addition to or in place of the existing automatically generated tags for linking purposes.
AutoWiki? will then automatically create a reciprocally connected wiki page of the same name for each node or link. The top part of the page would be a group-editable space that says what the connected node or link is and does. The bottom part would be for discussion about it. This portion could also be group-editable blank space, or it could be something more structured with identity makers, date & time of postings, tools to add linked uploads and live hyperlinks, etc., and be editable only by the poster and the moderators. For non-technical communities i would tend to prefer the latter.
Within the wiki, use of a page name creates a link to the named page in the usual way. But the page name in the header is a link that takes you to the same-named node in the dialogue map.
Each page would also have an assortment of finders’ tools accessible from it. These could include automatically generated pairs of links, one to each node to which it is connected and one to the corresponding page; a pop-up index of all the pages on which that page/node name occurs; and a link to an automatically generated alphabetical list of all the pages and nodes in the system.
I am not sure if documentation and evidence should be attached to the attached to the map or the wiki page or both, though my leanings are to just the wiki. I’m not sure what you do with a page, or documentation attached to a page, if the node is deleted. Maybe a permanent bin for deleted structures?
As someone who is not a professional programmer, I may be kidding myself about how hard it would be to create such a system, but it seems to me that the bulk of the hard work has already been done in the existing open-source code for Compendium and most wikis. I suspect that Compendium should go through one more generation of improvement in its asynchronous updating capabilities before we start adding features that depend on those capabilities, but this is only a not-very-well-informed guess. And of course you would not need everything imentioned on the first pass. For example, the finders’ tools could be added later.
Sam tells us that the reverse application, from an arbitrary Wiki to a map of the wiki as a dialogue, alredy exists in Wiki:VisualTour. If this is a straightforward add-on as it appears to be, I wonder why more people do not use it now? It is true that the map generated automatically by reference to CamelCase terms or the equivalent is not as meaningful as an ArgumentMap, because it is not the result of MappingArguments and because wikis are usually not arguments, and may not even be dialogues. For example, the links in an ArgumentMap or dialogue map may themselves have content, such as implication or causation, which are not necessarily there in the automatically generated VisualTour? map. Are there tools that should be available on the map level to get the most use out of such a map, besides the node-to page links? Especially for large and complex wikis, there is a lot that one could do with the map itself, such as cluster analysis.
Most wiki already support manually generating something that looks exactly like that.
Certainly an automated system would be nice, but I don’t see it as essential for “version 0.1”.
Your basic idea is correct: All these things are simple.
Where I think you start to go astray, is in calculating how long it takes to perform these simple actions.
I’d caution against asking big and involved questions here, and proposing a lot of ideas, unless you are have a goal of promoting from MeatballWiki:GuestRole to MeatballWiki:VisitorRole, or perhaps eventually MeatballWiki:CommunityMember role. (See the nice chart on: MeatballWiki:CommunityRoles.)
Answering the questions you are asking, by text, involves a great deal of back-and-forth, a great deal of text, and a lot of time.
The likelihood that we get to the questions that you really care about gets that much smaller.
That, and: It has taken me 3 hours to write what I have just said here.
There have been numerous rewrites.
I agree with Sam: I think the thing you want to do, is to:
One more thing:
Compendium is not OpenSource.
See: the Compendium Licensing Agreement, and then see the Open Source definition.
I am not a lawyer, but here is my interpretation:
I may be wrong, but I’m guessing that they think they’re going to sell a pro edition at some point, closed source, or they just want to keep that option on the table. Regardless, as it stands, they definitely don’t want anybody to take this software and release it in a way that they can’t control.
So, despite source availability, it is not what the OpenSource community calls: “Open Source.”
Dear Lion--
In asking about how simple or complicated it would be to link together an open-source wiki and Compendium in the way I suggest, what I am really trying to do is put it in a procedural category.
I’ve been on the Compendium Yahoo group for a couple of weeks. I actually posted the AutoWiki? idea there first. I have gotten a lot of enthusiasm for the idea from my fellow users, including one person so taken with the project that he volunteered to give some time to it. I have gotten two nibbles from folks with technical capability, but so far, they are just nibbles. I’ll be trying to cultivate them, and keep you posted. I only just learned that there is an entirely separate developers group, and i might poke my head in there too, although it looks to me like all of them are on the group I am already on as well.
One of the reasons that I put the AutoWiki? vision up here is that i thought there might be some folks on this list familiar with the inner guts of open-source wikis. It seemed like it might be at least possible that if i could get a someone who has been doing Compendium development on the same phone line with somebody who knows wikis really well, they could work out some really simple way of doing something that resembles what i was suggesting. Compendium claims that it is designed to be interoperable with other software. I quote, with parrot-like accuracy and degree of understanding, language from the Compendium website:
“If you’re technical, you can exploit our XML scheme, the Derby or MySQL? relational database, and public Java classes to connect Compendium to other databases and computational services. (If that sentence meant anything to you, then check out our developer website!)”
I have also identified the little cluster of interesting connections. One of the developers of Compendium is doing so as a consultant to the Knowledge Media Institute in service of a project called the Open Content Initiative, out of the Open University in the UK. The Open Content Initiative is a $10 million effort to put non-credit university-level courses and tools to self-organize or run classes with distributed learners on the internet for free use. It is funded by the Hewlett Foundation – which has also funded my organization, RP, and does considerable funding of sustainability work around the world.
(KMi is also working on a project called the KMi Semantic web portal that you might find interesting).
I have begun corresponding with Bob Horn (thanks to LionKimbro for pointing him out to me) about the particular problems associated with this dialogue, and we have scheduled a meeting in early January. Some of Bob’s more interesting public interest dialogue mapping has been funded by The Hewlett Foundation which has funded RP, and which is the primary source of fumd for the Energy Foundation, which in turn is the primary source of funding for me. As you can imagine, I find these connections quite intriguing, though i think it will be at least a year before i will be in any position to apply for grants for this work. Gotta get a non-profit set up first, for one thing.
This is another example of a lesson that i am learning only very slowly. For me, this project is entirely driven by the content. I have this big, hairy question, “How do you revive a genuine Left in the U.S.?” and nothing resembling a good answer on the horizon, and not even any very good method of proceeding toward an answer. And I don’t see anyone else who has a clue either. So I imagine this series of dialogues as a machine to generate and test candidate answers, and I start pitching it informally to acquaintances who are affiliated in one way or another with groups that will have to buy in on it if it is going to fly. A lot of them say “great idea,” but very few have offered to help. I mean, all those people are super-overcommitted. It is the nature of the business. And this is along-term project if ever there was – important, but not urgent, and so on the bottom of the stack.
But every time i have talked to people who’s central work is process rather than substance, I have gotten serious expressions of interest and offers of help. It is not a world i know very well, and i am still not sure best to use this fact, but i think it is going to be important. If people like Horn jump on board, then this could turn into something funded on the one-year time frame, rather than the four to five year time frame I had been anticipating. Having some hunk of funding pre-committed would certainly make me a lot less queasy about buying a one to two million dollar residential facility on a $70K income. (I still find it hard to comprehend that banks are willing to lend me that kind of money for a house purchase, but I have established that they are. Using a product called a “stated-income loan,” it’s not even a stretch, apparently.)
Oh, and i posted to the Compendium Institute group website to ask if they had plans for commercial development, or conversely, for full open-sourcing. I’ll let you know what I learn.
On a personal note, I have probably been spending too much time on this particular piece of this project, especially as i type with two fingers (that long ethics posting took me most of the night to type). Big, complicated and scary as significant software development seems to me, it is still a lot safer then some of the other key pieces, like spending my entire retirement savings on a 8 to 12 bedroom house in the hope that i can find people who want to get involved to fill it up, or cold-calling the directors of major non-profits to ask them to co-sponsor Compendium Dinners. One could argue that I have been using this part to avoid that part. But so far, it has been a lot of fun, and it seems to be opening some interesting doors, so I am not ready to back off quite yet. It has occurred to me that fancy custom software is a way to tweek people’s interest and give the impression that this is a serious operation, and for the next couple of years at least i will need all the help I can get in that direction. And I do seem to have a knack for gigantic, hopeless, decade-long projects operated on a shoestring, and some history of success with them – I and very much hope to have U.S. climate policy whipped by 2015 or so, and would like to have something interesting to take its place by then.
Here is the answer I got from the Compendium folks:
Re: [compendiuminstitute] Compendium license question Andrew, The current agreement is an artifact of Compendium’s institutional history. Our intent is to get to a true open source license. We’ve been working towards it for some time, but there are still some hoops to jump through. Al
Wow! Thank you for asking! That’s excellent news!
I’ve been wanting to write here; Just can’t do it this moment.
I am pleased to report that some combination of my prodding, fortuitous timing and a confluence of similar interests by other, has led to a lively discussion on the Compendium yahoo group on all three of the key issues discussed above: the proper architecture for to allow distributed groups to construct dialogue maps; several different approaches to linking maps to wikis, from simple imports of maps into wikis (TikiWiki is already set up to allow import of FreeMind? maps and for building such maps on the fly, though not, so far as i can see, for creating maps of the wiki itself) to full integration of Compendium with a wiki along the lines that I have described above; and finally, to issues around open-source licensing or lack thereof. It is clear that both some version of distributed construction and some kind of linkage to wiki-like capabilities are at or near the top of the development priority list for many people, perhaps including the developers who will actually do the work. See especially the threads: “Asking for what you want: AutoWiki?,” “Shared Compendium reqts and fanatasies,” “A user’s perspective,” and “Compendium license question.”
The technical discussions are well over my head, and i do not understand the implications of the architecture choices at issue. It would be great if somebody from here with more technical background then i have could stick there nose into this group and clue me in to what side i should be on. I have some related questions I hope someone here might have a view on, e.g. do we care with which wiki software they integrate?
I am trying to figure out what this page should be. I think this decision should be based on four factors:
I only have good information on the first and second, and need feedback. Here are some things this page could turn into:
1. A log of my quest for StrategicDialogSupportSoftware tailored to my project and needs, with occasional requests for support and comments and information contributed by others. This is more or less what it is now.
2. A description of my conclusions about StrategicDialogSupportSoftware tailored to my project and needs, with history and false starts edited out.
3. A catalogue of different kinds of StrategicDialogSupportSoftware, with lists, reviews, and links. I did some of this up front, but am doing lees as I hone in on my likely choices.
4. A discussion of what StrategicDialogSupportSoftware out to be, in the abstract and based on the community’s experience.
(1) is more or less what it is now. It is also the least work for me per unit of value to me. However, it feels a little narcissistic at times, and I wonder if I am giving anything to the community of value comparable to the help I am requesting from it, and occasionally receiving. (2) would also be easy to do. The stuff I would be editing out and tossing would be dross from my point of view, and the resulting page would be shorter, clearer, and closer to my model of a wiki page usually looks like. But for someone working on a project different from mine, the stuff I would be tossing, such as the discussions of the software I do not end up using, might be more valuable than what is left. (3) Seems more like the kind of page I imagine this wiki was designed to produce. It would be a considerable amount of work to create. I could take what I have and push it in that direction, but I do not think I would be likely to take it all the way to where it should be in this vision without consistent interest and participation by others. (4) would be interesting and stimulating, but is probably beyond my competence, though I suppose I could make a stab for others to improve. (4) might be regarded as an accompaniment of (3), such as an overview.
What do others think?
I was able to download and play with compendium. Some friends and myself even used it to try and map out a brainstorming session. Results are that I am pretty impressed with Compendium. So, it’s definitely good that they intend to create a more open “open source” license.
I think it is going to depend on how exactly Compendium folks decide to integrate wiki software, on the question of “do we care which wiki software they integrate?”
I am looking over the Yahoogroup list right now, and I’ll try to come back soon and talk some more here and there.
About your VisualTour? question above (sorry, I seemed to have overlooked it previously), each node in the visual tour graph was a wiki page. It was a script that told a GraphViz? server to basically look ahead as it went through pages, then draw a graph of all of the pages that were linked together. I don’t know exactly how they would do this in Compendium. I am going to guess that it will depend on what object is embedded in the page. In Wiki:VisualTour, the objects were CamelCase links to other pages in the wiki.in Wiki:VisualizeTheWiki, if you scroll down towards the bottom, they show the Graphviz source code, which apparently is looking at information from the web browser for what page you were last looking at. If your StrategicDialogSupportSoftware is not an application that is used through a web browser, I don’t think that this will work.
Lastly, maybe we can put together some conference calls to speed-up the brainstorming and discussion process, and out line those into this wiki (we can also record them)?
Dear Sam --
Your offer of participating in a conference call is greatly appreciated, and likewise for anyone else interested.
I have a news that makes it especially timely, and i may have to make a decision soon on which i could use this community’s advice. I was approached by a developer today, who may be interested in creating a proprietary version of exactly the software we are looking for in exchange for us helping with the requirements and conceptual development, and beta testing the result. They might even be able to help us raise some money for the underlying dialogue.
I have been becoming attached to the idea of doing this project mainly with open-source software, but I have to say this is tempting. They also mentioned the possibility of augmenting some open-source program but with additional proprietary add-ons, so they would be contributing to the value of the commons as well as taking from or competing with it, but this is by no means certain. I am not sure if i would most prefer talking to y’all first to clear up my own intentions or position, or after to help me assess the offer, if any, or maybe during, as full participants. The later seems to me to be ideal if there is anyone here who would like to be a full participant in the development process if we go the open-source route, and maybe less so otherwise.
The developers described in the last post want to meet with me early next week!!
The folks involved say they are inviting me to join the “a bay area initiative to create the next generation of visualization software.” Friends in the software industry are writing me vague but dire warnings to be careful and go slow, and get non-disclosure agreements and such, none of which i am currently inclined to do, but it sure has me unsettled.
I would like to go into this meeting as well-prepared as possible. Would anyone who thinks they have stuff to say that i should know about, concerning either the technical or the business aspects, please contact me? Talking on the phone would be great – maybe we could have Sam’s proposed conference call by then?
The folks in question are John Garn, director and founder of ViewCraft, and Jean-Michel Decombe. John looks like someone with a history of public interest and non-profit work and my impression of him is positive, though based on only a single conversation and few minutes of Google research. I know even less about Jean-Michel, and have not yet spoken to him, but he is a founder and director of Groxis, which is doing interesting work on what i take to be tools to visualize an arbitrarily-defined space of web-pages. Does anyone know anything real about either of them, or have suggestions on what i should be looking for??
I don’t know the exact situation, so I can only give general advice.
Immunize against this “next generation” hype. New software offers new interesting functions, hopefully. That’s it. Anything that can be invented in the IT industry now will be an old hat 5 years from now.
The question is whether the new features will solve your application problem. Probably not, because your application needs are a moving target. You can’t specify what is needed up-front. It is not visualization primarily. So you need a continuous, step-by-step development process according to your needs during the next 5 years. You want to build a community, so it must be community-oriented.
Wiki:ExtremeProgramming together with an OS base project could let you reach your goals whatever they are, given that you spend your money carefully, moving always in the direction of the community priorities.
So this is where I am so far:
I will need to decide whether try to go down the path of working with a commercial developer or working with primarily open-source software, perhaps as soon as next week. On deciding, I need to commit full-out.
If I go with the developers, I should form a true partnership with them, in which my use of their software becomes a highly visible and attractive feature of the Institute, one of the key things that makes us stand out. We would not merely beta-test the software: we create buzz, among some of the most influential, interesting, and well-connected individuals in contemporary politics (at least in the non-profit sector) for a product that will ultimately replace the Group (e.g. the Yahoo group) and the wiki as the most widely-used web communications tool for distributed consensus-building, decision-making, and collective action. The goal is to contribute to the design of a product that would become the industry standard, purchased by perhaps five or ten percent of the more than a million nonprofits in the U.S., plus more abroad, and a reasonable share of the for-profit enterprises as well; that will make those enterprises more effective and so contribute to a better world. The developers in turn have a stake in the success of our dialogues that causes them to cheerfully recruit large donors to finance it and big-name entrepreneurs to endorse and perhaps participate in the process.
If, on the other hand, I go with the open-source folks, our commitment to open-source development of software as a common asset shared by all becomes a part of the Institute’s identity and public persona. Software and music, literature and the genome, the formulae for AIDS drugs and for Coca Cola, all are part of the universal heritage of humankind that ought properly be made available freely as a birthright to all. Our failure to do so is one of the most basic irrationalities produced by our current economic order, where people sweat and suffer and sometimes die trying to acquire knowledge that may be costly to produce but is essentially costless to reproduce. We will call on the open-source community to support our efforts and do our best to learn the lessons of that community in how to create sophisticated, productive, output-oriented distributed dialogues around difficult and complex topics. We will also seek to understand the business model of open-source enterprises create analogous processes to support our activities where possible.
On the developer side: Plus
Minus:
On the open-source side: Plus We are philosophically and politically aligned with open-source development.
Minus:
I am still trying to specify and clarify the criteria that I will use to answer my original question – how to decide which rout to persue. This is the best I can do so far:
1. How fast can the developers get a product that draws a map, and publishes it to the web, and supports links in and out up and running? Are we agreed on the priority of the core set of mapping and wiki/forum functions that I outline above?
2. Given that we are contributing valuable services not only to the development process, but also to creating buzz around the product, to providing non-advertising advertising that is carefully targeted to opinion leaders and influential individuals, it should be in the interest of the developers to help make the dialogue process fly. What resources of finance and prestige can they bring to support that effort? Could they supply $50k/year for two years, with some of that being framed as matching money to challenge foundation grants, or would it be overreaching to ask for that much?
3. Is Compendium development fast enough to give us the capacities we need on the timetable we need? Can we find any developers in the open-source community who are interested enough in our project to give time and effort to pushing Compendium in the directions we need?
4. Is there a “middle way,” in which the developers augment the open-source software by adding better and more transparent linkages and then construct proprietary add-ons, making their money doing training, customization, and by adding new functionalities for hire.
A language issue: both ways you have to do with developers, either (1) “developers of proprietory software” or (2) “developers of open source software”.
There are some differences worth noting:
This puts you in a paradoxical situation with no “silver bullet”.
If you are able to put $50K or more into development I think that (2) is the better choice. It puts you in control of the process and you can produce a value for the community. I guess that a 3-year times $20K development process would give you a 50% chance to succeed, a 5-year times $50K would practically guarantee success (90%), more money would probably reduce the chances again.
If you want to go without spending money, I’d suggest (1). My best guess is a chance of success of 30% within 2 years. It will probably fail because the company has to go not towards your needs but towards hype-able features for the enterprise market.
I don’t know if this was already mentioned but companies like SocialText and JotSpot (both in the ‘enterprise’ wiki market) would also be very interested in getting some marketing buzz. Additionally the SocialText product is OpenSource.
I liked this quotation from Socialtext CEO, Ross Mayfield: “Simple tools with simple rules yield the best results.”
I admit that i would be more inclined to go open source if I had any offers from open-source folks to do development for us – or even a good way of finding such people. Again, this project has no budget at this point, except what i take out of my retirement savings to finance it – hence the appeal of seeking a contribution from for-profit developers. We are trying to find some way around the chicken-and-egg problem to create momentum.
Just a link to International Association for Public Participation an organisation with similar goals?
I think there are many ways to contact OS developers. There is the Free Software Foundation that is politically orietned and might take interest in the project. Nominally a million developers use the sourceforge and other open source platforms. Here in Austria e. g. we have an organisation OSalliance http://www.osalliance.com/ which organizes OS developers and service providers for larger projects.
There are also great people like HansWobbe who have enough money to give any support to a project they are really convinced of. Talk to him, I think you could be on the same wave-length.
I’ve been hanging low; Quick thoughts:
This may be a project where a form of PinkoMarketing/ SocialSynergy:PeerInvest, or similar-type fundraising could make possible a fully funded OpenSource project. This is one of the major goals of SocialSynergy:PeerInvest, and of my involvement in BarCampBank: to create realistic models for community/peer funded projects.
In this case, I envision that it would be possible to first envision the product of development as much as possible, and then seek out and/or attract the different people who can create/program the system into creation.
Attracting investment would be best done by creating an at least partially “open” business model, which lays out the general direction, and intended markets, and possible markets that can be generated, by the existence of the technology and the knowledge of how to use it. One interesting way to pursue this is through a for-profit cooperative model, where owning shares in the cooperative=sharing in the long term revenue generated. So, investors share in the revenue, developers share in the revenue, etc etc. The cooperative could also serve to foster the business entities that create business models around the technology. This is the type of thing that we are discussing with BarCampBank.
Sam, I really want to talk to you about this. If you are willing, could you email me your number? I’m at hoerner(at)redefiningprogress.org
Here are some thoughts out of left field:
The reason I ask this is because:
For instance, if you have a notebook, with 80 pages, you will find: “Wow! I can find everything I need in here!” But that’s just because the use is small. As you stretch out to 500 hundred pages, you find that you can’t find anything anymore: Scale matters a lot when you’re talking about an IntelligenceSystem?.
Compendium is suspicious to me, because it’s high on structure.
I have a favorite paper in this territory: It’s called: Formality Considered Harmful: Experiences, Emerging Themes, and Directions.
Briefly: Conversations and problem solving operate in a sort of hysteresis: They don’t tend to flow in water-fall “A → B → C” fashion. (See: section Opportunity Driven Problem Solving in Wicked Problems & Social Complexity.) Rather, they jump all over the place. Nor do our expressions fit a nice even water-fall pattern. Nor do rules in a system. (Why we favor soft process, rather than hard, here on CommunityWiki and MeatballWiki.)
You asked the question once here: “If someone shouldn’t be allowed to do something, why doesn’t the software here account for that?” The answer to that question is directly related to the “Formality Considered Harmful” paper. It takes far too long, and requires much too much explicit social agreement, and so on, to express in software the rules of the social actors. You have to teach the system what people are, then what their security priviledges are, and then you have to write the software to either permit or to deny. And then by the time you’re there, you don’t want to change anything, because now someone has to go back and change all that stuff. (That would be AlexSchroeder.) And social rules change all the time. I mean, you are seeing it happen now, with the CommunityWikiGovernment.
So too it is in discussion, with presenting points and counter-points and so on.
Consider my post in EthicsDiscussionB, near the top. How would that have looked, in Compendium?
It would not be the same thing. It would not be the same thing at all. The communication I made would be almost impossible in Compendium. But the spectrum of communication in regular human dialog is incredibly diverse. We exhibit here only one tiny little part of the spectrum of communications.
And yet, we know, that it is useful to see things arranged, and formalized. What is the answer to this?
What I think, is that:
systems that emerge.That is, I believe in incremental notekeeping and manual reworking; Not trying to get things into an ideal representational form at the beginning.
This brings us straight to the LackOfReworking problem, which is just as much a problem in the material world, as it is here in our virtual wiki world.
But I believe we have found the answer to the LackOfReworking problem: Make it somebody’s job to do it.
So, in summary:
Tools like Compendium may help the person make the chart and diagram that you want in the end. But I would want to test it for 3 months, with a group of people, before I committed to actually hosting full-on discussions on it.
This suggestion is actually quite in tune with my own gropings. I have been urging the Compendium folks to create a version that automatically creates a wiki page for each node or link (but not conversely. I think most of the discussion will happen in the wiki, while the map will serve as a sot of visual index and overview. The initial map will be done by us, from the raw material of some in-person conversations. But i would still like participants to be able to edit the map, even if they don’t do it very often. I think it is important to a sense of ownership of the larger dialogue to which the individual dialogues sum. More on structure below.
I have not given the software even a minimally adequate test. However, i have located some people who have used it in a business environment, and am arranging to interview them. Pointed question suggestions welcome.
Actually, I do have in mind a software program that would be:
I believe that InkScape can be adjusted, in relatively minor ways, to be this medium.
First, you fork InkScape.
Then, you make a few changes to it:
If you did all this, which is hard, but not too hard, you would have a visual, open ended, collaborative, OpenSource platform for communications.
If you want to do Compendium style stuff, it would be easy to do so: You just reshape the text however you like (squish it up real small, or expand it to be big, and so on,) put down whatever icon you like, and so on.
Drawbacks:
Now, this project I described, easily fits Helmut’s timeline. (Which, is more or less accurate, by my view.)
Yes, quite right.
If I were to do this, here is how I would do it:
If the project is really succesful, and lots of people start using it, there are ways to spin businesses out of this: If you can convince some large company that this is a great way to do documentation, host discussions, and so on, then you can do the SocialText thing, and offer training support, feature addition, integration, migration, and so on. RedHat? does this, SocialText does this, and so on. People make fun of this, because it’s kind of a new way of doing things, but it really works, companies are making money (more than they spend) doing this.
Then again, all this stuff we’re talking about: It’s aweful experimental, aint it?
Why strive for the next generation of tools, when most companies aren’t even up to speed on the “present” generation of tools?
What I mean is:
Most companies are using something like:
I’ll call that “previous generation,” where “previous” means “2000.”
I’ll make up something and call it “the present generation,” as in 2005.
It has such rocking amazing advances such as:
If you wanted to really rock the boat to 2006, you could introduce things like InkScape real-time, Gobby real-time, but since y’all will be living together, it’s not so important.
Believe it or not, in my experience, most companies are still in 2000, with only a few shoots of 2005 coming up.
Platform 2005 has some maturity now, and can be wired together without too much work.
To aim for Platform 2007, though, might be a little risky?
And I mean: Your goal is to have your conversations and answer your questions, yes?
Do you also want to simultaneously be testing what we call a “bleeding edge” platform, at the same time?
When I say, “bleeding” edge, I mean: “There’s blood dripping down the rough hewn blunt metal, because people have been cutting up their internal organs on the deformaties of the thing.” It’s called bleeding edge because its so new, it’s broken.
On CW, we work on some bleeding edges, because, well, that’s what we’re here for. We’re here for that.
But, are the people in the community you are working with going to be there for the comm. tech? I’m a little skeptical, …
You might want to consider platform 2005.
And then, take baby steps to 2007: Cheaply solve the ContentRouting problem. I’ve been meaning to rewrite InkscapeToOddmuse for a while now, anyways; I wouldn’t need to charge anything, it’s easy enough to do in my free time. I’ve been wanting to make it work to a generic “Image Server” for a while; that’s a project I nurse in my head.
Dear Lion –
As always, I find your posts stimulating. Moreover, as seems to be becoming the pattern, I seem to be stimulated mainly to disagree. I am not quite sure what to make of this, except to say that I find it more fun to disagree with someone when I am not at all sure that I am right. And unlike ethics, where I would say we are fairly matched, here I am clearly making claims in areas belonging more to your field of expertise than my own. Still, here goes. Do correct my errors, and note that i responded above to some of your other comments.
There are many, many cases where you want to bring a group of people together to engage in a discussion of make a decision where the commitment of the people involved is limited and the organizers do not have direct authority over the participants nor the resources to compensate (or threaten) them. This is the typical situation in collaborative processes in the NGO setting, where I work, and a common one in academia, user groups, and many other settings. I think of it as the problem setting for the software we hope will be the solution.
In such a setting it is essential that the cost of entry be low and the rewards immediate. Organizers of this sort of process may have significant investment in making them go well, and can be asked to spend some time climbing the learning curve. Also, once a participant has bought into a process and become invested in its methods and outcome the participants are often willing to devote genuine resources to making it work. But before such buy-in takes place, even quite minor barriers will cause potentially valuable participants to put their energy elsewhere.
Something I would put in the category, not of a minor barrier, but of a significant one, would be the requirement that participants install a piece of software and learn how to use it. For most of the on-line dialogues I deal with on a daily basis, this would be enough to guarantee that the process collapses before it gets properly started. So I think you just have to have a browser-based interface for any process that will not be done by the moderators or a committed core. If that means you lose interesting functionality, then it does.
I would also add that, while the degree of flexibility that you are suggesting is undoubtedly and spectacularly cool, also wonder if it would really serve the particular purpose I have described well. It seems to me that the mandatory objects-and-arrows character of most of the mapping software I have looked at constitutes the grammar of a sort of visual language. Eliminate the grammar, and you can make a more expressive range of pictures, but you lose the meanings embedded in the grammar. I kind of see this collective whiteboard + text as a cool thing, that might well be useful to the overall dialogue process I describe, but not as the mapping tool itself. (I am still leaning toward Compendium for that, but – and this is the topic for another post – it is bothering me more and more that Compendium maps are ugly. You sort of don’t want to spend much time studying a Compendium map. Some other dialogue mapping programs, though much less full-featured, have nonetheless got much better esthetics.)
It is worth noting that people generally feel like they know what a browser is and what it does from experience, even though they lack any technical understanding of the process. Moreover, I have observed that this familiarity can result in people feeling like they are using software that they know how to use, even where the procedure that they must use on a remote page is quite different from what they most often do and has complex and subtle coding embedded in it. This is especially true where the actions can be identified with common actions that people take on web pages, such as making a choice by point-and-click, filling out a form, etc. So I point and click, and suddenly I am playing Bejeweled, which is not at all the same as the usual radio buttons, or a fill out a form and get back instructions on how to knit a sweater. The browser interface tricks you into thinking you already know things when you don’t.
Most people in my circles know how to use Yahoo Groups and the like (though even there some are challenged by, e.g. the file storage functions), and blogs; but only a few have ever tried to contribute to a wiki, and hardly any have used any graphical representation software outside of MS Office. So to respond to your bleeding edge question: for my personal project, I want to have something that works in three months, four tops. It does not need to do everything that it ultimately will, but it has to have a way for naive people to get in immediately, and it has to display the vision of what it will become, either immediately or shortly thereafter. I think it is fine if it displays it by having us do by hand things we ultimately want the software to do.
Beyond that, the question is whether it serves the dialogue or detracts from it. Do shiny bells and whistles that nobody in this particular group will use attract money to develop more basic features and finance the underlying dialogue? Then let’s pour them on, by all means. Do I end up with a great piece of software and a bunch of individual dialogs that never go anywhere because they are never connected, because the tool that we were hoping would connect them got hijacked to an internal development dynamic of some kind, whether open-source or commercial? Let’s not do that. I’d rather just have a Yahoo group. Can we do something innovative and interesting that also genuinely contributes to the functioning of the group? That is the best of both worlds, and would be fabulous.
The design challenge I throw out is this: It has to be easy to use without training. Assume that no feature that un unsophisticated user will require more than 2 minutes to figure out will ever be used. The formatting features here at this wiki, for example, are far too complex for this group. If you have to search for the page that contains the instructions, you’re dead. Better to copy the much less elegant set of editing buttons from the top of MS-Word, and let people believe that they already know how it works.
More generally, it will not work to do the easy things badly in order to do harder things well. To take a not-so-random example: I never did get my picture up right here. I struggled with picture-editing software that I hardly ever use for 45 minutes to try to do reasonable cropping and resizing to a pixel-based standard (which I had never done before). Then I could never did get the insert instructions straight, and somebody else had to do it for me. In my setting, pictures are critical, and there are lots of people whose voices we will need who are way less technical than I am. I’m an academic, and I’ve written macroeconomic models in C++. When I am trying to get somebody my age who as spent the last 30 years organizing Farmworkers to participate, I really want the site to be able to resize his picture for him.
Lion, your ideas for improvement to InkScape are fantastic! I’d love to see some of those changes you suggest above. I agree with you that people are often in the area of about 1998-2000 when it comes to software and hardware technology they are using.
Andrew, having helped many, many people learn to use SocialSoftware over the years as part of my business, I can see where you are coming from as far as ease of use concerns go. This is often one of the biggest hurdles successfully implementing many SocialSoftware systems. Especially with people who do not easily uptake new evolutions in social software.
Yet, I can tell you from direect experience that the SocialSoftware platform that you create and use does not have to be totally geared towards the unsophisticated user in order for you to succeed, and in order for it to be used by the unsophisticated user. After all, these unsophisticated users are out there buying complicated technology, like surround sound DVD players, and programmable coffee machines, and who knows what else. So, they can deal with complex technology, and they can read manyuals. The problem often is, that there is no “manual” that can detect and show all of the individual frustrations that people are out there strugglig with.
I have observed, in working with people and showing them how to use software, that often people know more than they think or say that they do about the technology they are using. People will say “I’m not very computer savvy, I don’t know much about this”, yet, I often find they know almost a huge portion of they need to know about what they are trying to learn. I like your example of the web browser, and how say that people know what a browser is and what it do