User:Eleanorg/Journal 2.2

5 Feb

Clearing up my brain to know what to do next.

Project history:

• Dissolute Image: a file only published with the endorsement of many participants.
• How could this logic be applied to the problem of how to curate Radical X content?
• Idea: a magazine where each article needed to be 'hosted' elsewhere

• Instead of starting writing code for this, stopped to pause & reflect on 'gestures of endorsement' - could there be others?
• 2.1 - Residency: transmitting repetition; what is relationship of person transmitting to the content they transmit?
• 2.2 - Consent. Reading up on feminist theories of consent: signifiers are often unclear. 'yes' is transmitted but is it meant?

• Project plan: an ongoing discussion about consent, ending in a publication. Supported by editing software that visualizes the extent to which people endorse the content.
• Experiments
  • Modular ideal of consent: Endorsing individual sentences - Mediawiki (not best tool)
  • Visualizing level of endorsement through typography.
  • But consent must be encoded in the interface: Researching how consent/agreement encoded in online interfaces.

• Current possibilities:
  • A 'voting' system to visualize popularity using graphic design. (eg [| Slogan Voting].
    • Problem: populist approach; doesn't address fact that voting systems force decisions, perform in their own right
  • Interface/s to encode consent [Interactivos proposal]
    • By contrasting the options offered by different systems, the way they 'manufacture consent'/ create different output can be critiqued.
    • Problem: if only one is made, problematic implication that this system accurately encodes meaningful consent.
  • Original idea of individuals hosting own content, magazine aggregating it.
    • Problem: populist approach. 'Problem of silence'.

Dilemmas:

1. how to create a system which is consensual (democratic), but not simply 'populist' (most popular content wins)?
2. How to be critical of the way that interfaces encode consent, thus perform in their own right?

31 Jan

Idea for workshops: use workshops to generate discussion & new slogans. Then put these into the software.

To prototype next: use 'Bang with Friends' idea as editing tool. Invisible until both agree.

30 Jan

Working on the visual design of consent slogan voting system. Thinking about changing it from css to SVG - easier to have control over layout, and potentially easier to print to .PDF as well. TODO next: look over code from yr1 for generating svgs and copy paste.

by the way... Just found this on the wiki. Possible basis for future project. Perfect combination of elements: painstaking collective labour, gender, human computing: The word computer was originally used (circa 1900's) to describe a person whose job it was to compute.

http://www.cfa.harvard.edu/~jshaw/pick.html

"Pickerings Harem" was a group of women computer's who translated astronomical data recorded on photographic films into tables of data in standardized units.

25 Jan

hello journal. Today it's How I Learned to Stop Worrying and Love Data Visualization.

Now thinking for the forseable prototyping future of my project as a data-viz one, not only ETHICS but AESTHETICS of consensus or lack thereof. Re-visting a lot of code & ideas from trim 2.1 and 1st year, and my interest in translating numbers into visual things. I think this is what I enjoy most about writing code actually. So I'm going to do that for a bit, and enjoy it, and make some things and then we can have the serious conversations later. OK!

BUT. In the meantime - note that current version ADDS each vote to the previous total. So if someone votes 'no' (0), it stays the same rather than having its total reduced. This mirrors the 'common ground' exercise where people take a step into the circle when they agree with a statement - everyone ends up standing together, because disagreement is ignored by the system while agreement gradually accumulates. A more sophisticated system could deduct votes. Although this could lead to edit wars. I like the current model, where things get bigger as they are consented to - the other ones fade into the background.

It could also work to add annotations - so people cast their vote, but written feedback is also stored and can be accessed for example on a hover event or summat.

Interview about a book about collaborative media in Prototyping Futures: http://medea.mah.se/2012/12/collaborative-media-production-consumption-and-design/

24 Jan

50-pack tattoo consent forms.
Somehow arrested by this image/marketing of this 50-pack. Industrial-scale anticipated consent/ encoded consent/ consent as business --- link to callcentre work and its relentless ruthless search for 'consenting' subjects.

24 Jan

I'm stuck on what to make next. Was hoping SMWiki would offer a way to tag and sort texts and it didn't really. I could start writing something that does, but now I'm wondering - do I want to make yet another tool for annotating online texts? What is my position if I do this? Am I claiming that this system is more consensual? Am I claiming that it accurately records a process of social negotiation?

I'm coming up against the problem that these systems always perform in their own right; they won't ever 'record' an analogue conversational process. They make its results accessible, but in doing so alter/produce those results according to its own logic. Eg - e-consensus constructs consent as a 'green light', something given. Wikipedia constructs consent as a default in the absence or conflict, or else a temporary agreement following debate. Do I want to go down the road of looking at how we encode consent, ie how our ability to 'give' consent is constrained by the gestures available to us? Returning to the essay The Nature of Consent. I found it useful how it differentiates between three key aspects of consent. I think I need to clarify which one/s I'm tackling:

• The grammar of consent

Challenging the problematic grammar of 'giving' and 'receiving' consent (as opposed to constructing it collaboratively)

• The ontology of consent

Ie, is wanting the same as consenting? (consent ≠ desire)

• Signification

What gestures should be taken to signal consent and which gestures are invalid?

Feminist theorists tackle some or all of these three areas. The ones I'm most interested in, those arguing for consent as a collaboration, challenge the grammar of consent as well as the ontology of consent (consent should not merely be 'given' because consent ≠ desire; to close consent/desire gap collaboration is needed). Others such as Hugo Shawyzer examine the ontology of consent - 'yes (consent) doesn't always mean yes (desire)'.

Looking at software systems, the last one - signification - seems the most relevant. Ie, how is consent encoded. (Green button, reverting an edit... etc). And can it be encoded in such a way that challenges the traditional grammar? Wikipedia starts to do this, so does e-consensus albeit in a limited way.... But the better it models this notion of collaboration, the less software-like such a system becomes. So it seems silly to write software to mimic human conversation; software is useful precisely when you need to SIMPLIFY real conversation in an encoded way. So actually my interest is much more a graphic design/data-viz one than a software one. Not 'how to design a collaborative system' (because the answer is obvious - conversation). But rather: how to capture this conversation in the final output? How to challenge the idea, which props up hierarchical editorial power as well as coerced 'consensus', that the 'final outcome' must be neat & tidy. Indeed that there must BE a final outcome.

A system for a messy outcome. Challenging the assumption underlying the Wikipedia book maker: we want to print out something tidy; the consensus position. No: I want to print out something messy, something honest that reflects the struggle over its own creation. So I can use pre-existing systems. The bit that I want to mod is the PDF output bit. BOOK WITH THE SOUND OF ITS OWN MAKING

next: make something quick that visualizes version history. Ie - translates numbers (number of edits?) into visuals (font size).

23 Jan- thesis therapy w/ Dr Steve

• Use same methods in thesis as in practice: exploratory, empirical, conversational
• WHY don't I want to construct rational/scientific argument? A: 'Importing' out-of-place ideas from one context to another gives fresh perspectives on an question. Also personally - an attempt to negotiate different registers, see if it's possible to speak across registers/discourses. Also politically? - infuse feminist ethics into collaborative production.
• Q: Who am I speaking to?
• The topic is genuinely problematic. The dilemma is productive though.
• ALLOW my subjectivity in this dilemma to be an instrument for articulation. It can be convincing; it becomes rhetorical - don't pretend to be neutral.
• Link between 'yes means yes' & consensus/wikipedia: social codes produce behaviour, whether they are slogans or software. Encoded either socially and/or symbolically.
• Describe clearly how these different domains map onto one another. Eg, where are feminist definitions of consent relevant to software design?
• Include TM workshop and other experiments.
• See 'Us', teams painting over each other's work. << Seems to make a point about human nature but in fact sets up a scenario to produce conflict and trashed paintings.

23 Jan

Just came across this random rant on wikipedia about 'e-consensus'. Fascinating; it says "The e-consensus paradigm views the "e-document" as a "battleground of ideas" for wisdom and ignorance to forge a common decision of what is to be said and how." And it has been listed on Wikipedia:Votes for Deletion. How perfect! http://en.wikipedia.org/wiki/User:Quickwik/E-consensus

Playing with the fascinating https://econsensus.org/ Trying to encode discussion, comments, warnings & ultimately consent. You give consent by pressing a green button...

This from the website of Activate, the NGO who developed the software..

   Participation and Consent

People have a right to be heard on matters that affect their wellbeing - so we are morally obligated to listen, and help others raise their voice. No development project can effectively serve its recipients (or its users) without their frequent input during design and early testing - so we are practically required to get our users and beneficiaries involved early and often. This drives participatory international development and agile software development. It also drives our company's governance and management. We are committed to ensuring that everyone in our workplace has a voice in policies that affect them. We accomplish that by using participatory decision-making and testing each policy to ensure it won't prevent us from serving our clients or endusers and it won't prevent any of us from working effectively in the company. It turns out that this reduces our risks, helps us learn rapidly as an organisation, and increases our sense of satisfaction and engagement: doing the right thing is the most practical choice as well.

team painting as corporate teambuilding

check out this use of collaborative labour as teambuilding awayday: http://www.platepus.com/people/painting.php

And peer-review via Github: https://github.com/cwcon/push

21 Jan

Notes on Wikipedia's book generator:

• You can't add history pages to the book, but you can add talk pages.
• Table of contents, Index are generated automatically and it looks pretty.

I think I want to propose a way of linking the 'talk'/history of a page meaningfully to the text - not just separate 'talk'/history pages.

Notes from Michael tutorial:

• Shortcoming of RDF - data is removed from its original context; 'facts' appear to stand alone & reified in triple format.
  • RDFS attempts to put 'facts' inside HTML docs
  • RDFA vs MicroData - two projects for adding structured info/annotations to a page. MicroData - getting DOM access to the page to grab specific bits.
• Check out Ted Nelson on Hypertext - early ideas on semantic texts. See his Filesystem for the Complex... essay on the need "to provide

the capacity for intricate and idiosyncratic arrangements,total modifiability, undecided alternatives..."

• Wikipedia 'Neutral Point of View' policy masks political realities. (Also see WikiData presentation, which makes a nod to diversity of data sources/definitions.)

• Tools to visualize version history/make annotations:
  • Brainch by Stdin: http://stdin.fr/Works/Brainch
  • Epicpedia: http://www.epicpedia.org/ extract history for 'visualisations'/re-use elsewhere.
  • ShiftSpace by Michael van Schaik & others: http://www.shual.com/shiftspace/
  • DeptOfReading: http://automatist.net/deptofreading/wiki/ Simple html-based way of 'annotating' sections by searching for them when quoted elsewhere & wrapping in a CSS class.
  • Check out javascript libraries for working with selected text.

• Alternative proposals for wiki-based collaboration:
  • Liquid Threads mediaWiki extension, to replace Talk page with threaded forum.
  • See also Smallest Federated Wiki from Ward Cunningham; everyone has their own wiki & pushes/pulls to/from others - videos at wardcunningham.github.com
  • Oral citations: critique of current citation policy (funded by WikiMedia Foundation grant): http://meta.wikimedia.org/wiki/Research:Oral_Citations

• Good not to make yet another 'solution' to problem of semantic annotation. Projects above don't deal so much with difficult social question of what constitutes agreement.

18 Jan

Trying out Semantic MediaWiki as a way of collecting and tagging texts. Coming up against its limitations is helping me figure out what I /do/ want. Notes:

• It's actually easier to copy & paste fragments of text than to tag them w/SMW syntax
• I want to copy out portions & write next to them, not just mark them up. (Possibility: use Page datatype as default and put comments on the tagged phrase in the empty page which is automatically linked to when tagging.)
• Need to make inline comments, not just in the talk page. When I make a note (eg 'this is bollocks') I want to say why - like a log message.
• There needs to be multiple layers of property tags - one for each user who has tagged within a page - so that one text can have the same sentence can potentially be tagged with two different properties, by different users. It's not possible in SMW to have tags overlapping or nested.
• Workshop idea - group decides what ontology to use at the start.
• Maybe this is too meta, but the flat structure of wiki & flexible property tagging make the issue of coordinating (ie agreeing upon) the ontology vital. Would be massive shame to pre-define an ontology (eg property names) as this avoids a problematic democratic process.

• Clear need for a collection of texts as a precondition of the project, for any group to work with. Start this sooner; wiki is a good way of doing this.
• Want to keep participants' comments separate from the texts themselves. So that it's clear what we're reaching consensus on - not a sprawling open-ended debate. Introducing limitation forces issue of consent/sus to arise.

16 Jan

Thesis time! Terror! Headspace...So. Narrowed down my technical focus to consensus processes for collaborative document editing. Have excluded earlier interests in aggregation, hosting etc in an attempt to avoid the tunnel of doom. Michael convinced me that spending some time prototyping with wikis would be a good idea. Their functionality plus the history/practice of their communities gives a nice correlation of theory + practice.

SO. My topic of thesistic exploration will be the tricky border where 'consent' meets 'consensus'. ('consent' = agreement of an individual; 'consensus' = agreement of the group.) The aim is to critique the way that consent is encoded in:

1. Dominant consensus practices, and thus
2. Collaboration systems encoding these practices (Eg wikipedia).

The critique will be made using feminist theories of consent. I will argue that:

1. The rhetoric/ideal of consensus often masks the ambivalence and compromise experienced by the individuals 'giving' consent.
2. Accepting this ambivalence/not-consensus is ok and beautiful and a powerful ethical position.

Here is the outline, with research objects in << >> brackets:

• Introduction
  • Demise of hierarchical editorial power online raises question of how to edit collaboratively.
  • 'Consensus' is a popular model used by projects like Wikipedia et al, encoding the popular activist approach to organizing.
  • Question: how consensual are these consensus systems? What is lost in the quest for consensus and who loses out?
  • Argument: interrogating consent at a low level gives tools for evaluating how consensual 'consensus' process are. Feminists give us these tools.

• Summary of the concepts of 'consent' as opposed to 'consensus'.
  • Trace the the source of 'consensus' process in Western leftwing projects
    • Q: What common sources do consensus evangelists (Wikipedia, Seeds for Change, Movement of Movements etc) draw from? << Historic texts/groups; Seeds for Change interview >>
    • Q: What assumptions does this concept of consensus make about individual consent?
  • Summarize feminist campaigns on consent << Reclaim The Night; Rape Crisis policy >>
    • Argument: interrogating consent at a low level gives tools for evaluating how consensual 'consensus' process are. Feminists give us these tools.
    • Summarize origin & dominance of the 'yes means yes/no means no' slogan. << Historic texts/ campaigns; RTN interview >>
    • Outline correlation of this slogan with the legal definition of consent. << 'The Nature of Consent'; 'Consent to Sexual Relations' >>
    • Introduce contemporary debates questioning this slogan. << 'Yes Means Yes'; Hugo S. >>

• Consensus in online publishing
  • Introduce the issue of how to curate crowd-sourced content. (DANGER of diversion into filter bubble debate)
    • Hierarchical editorial control being replaced by user curation / collaborative curation.
  • How consensus is encoded in collaborative editing << CHOOSE RL EXAMPLE/s. Possibilities: post-Indymedia projects; Wikipedia >>
    • Describe an example or two of how consensus is used in collaborative editing. (DANGER of becoming a Wikipedia/WikiMedia thesis)
    • Use feminist theories of consent to evaluate how consensual this process is.

• Proposals for consensual approach to consensus
  • Ambivalence is good
    • Back up with feminist/ direct democracy theory (Q: how/where to include theory on consensus democracy?)

potential fork alert

Writing this I realize that feminist theories of consent can be used powerfully to critique practices of consensus. (Ie, that they are not really consensual). BUT. They are two different topics, with different implications when it comes to software.

1. Consent: TOS, 'I agree' buttons
2. Consensus: wikipedia, comment rating systems.

The former is about individuals giving themselves over to a policy. The latter is about individuals expressing their views. They are related, but are they related enough to both be the topic of a project/thesis? I reckon........... the point of the thesis is to bring them together. BUT. It will become too sprawling if I include how individual consent is encoded in software ('i agree'). Instead I will limit the discussion to collaborative systems for deliberation. BECAUSE: they most closely model the idea of consent that feminists are proposing, albeit with more subtlety needed (which is what my graduation project attempts).

Solution: make the following analogies:

1. TOS/ 'I agree' = conservative status-quo of consent
2. wikipedia consensus = attempt at radical, truly collaborative consent

Thus maybe mention 1. briefly, but my main interest is in 2.

15 Jan

Have been converted to the wiki cult. May be gone some time.

See this for an example of the prized concept of consensus in FLOSS community: https://gnu.org/consensus "I'd like to reach consensus on officially supporting this manifesto" https://lists.gnu.org/archive/html/consensus/2012-12/msg00009.html

mediawiki experimental thoughts:

• love wikis. easy cut & paste BUT then requires rules if metadata is to be coherent.
• functionality lacking from mediawiki by default:

* keeping track of where a text has come from (see active archives approach)

* ability to browse tags by who added them (possible w/ semantic mediawiki?)

* ability to add multiple layers of tags to a single page w/out creating a new page, or overwriting page in the VCS

• q: how to integrate the version history with meaningful metadata?

C's observation was useful: I'm trying to make something halfway between a VCS and a metadata system.

9 Jan

Going to join Dave's browser editing club this month and do editors that use different schemes for negotiating consent. First prototype will be a text-editing sketch where a block of text can be reviewed by the user for publication. User can highlight the portions as 'yes' 'no' or 'maybe'; they are stored in a db with these phrases as tags. A separate script that outputs PDF will format the text for print such that these judgements are reflected in the design. (eg - 'no' is omitted, 'yes' printed, 'maybe' printed sometimes at random, or fainter.)

Need to find some js-type code that lets you 'tag' portions of text by highlighting them. C recommended CodeMirror - js that creates an in-browser editor. It comes with optional extras that add the type of functionality I want, like putting text into variables when it's highlighted by the user.

Next steps:

• read CodeMirror user manual.
• make a prototype editor that puts selected string into a variable.

from http://codemirror.net/doc/manual.html#event_cursorActivity Events

A CodeMirror instance emits a number of events, which allow client code to react to various situations. These are registered with the on method (and removed with the off method). These are the events that fire on the instance object. The name of the event is followed by the arguments that will be passed to the handler. The instance argument always refers to the editor instance .."cursorActivity" (instance)

   Will be fired when the cursor or selection moves, or any change is made to the editor content.

...Read user manual. Useful bit = detecting cursor events that mean text has been selected. Looking for a simpler way of doing this in js myself. Seems there is an inbuilt method, .getSelection(), that does just this. Find out how to use it. Try using the sample code at http://stackoverflow.com/questions/7416555/getselection

27 Dec 2012

My project has a social side and a technical side. I've put time into figuring out where to intervene socially. Now I need to firm up a software spec and learn some skillz.

FEATURES

• Anyone can contribute to the 'library' (or - limited to a group who've committed + have passwords?)

• Users curate by committing physical resources as a marker of agreement/solidarity (Piracy Project as opposed to Assembling Press, where authors must print own work)

• Popularity is visualized physically (see Amsterdam Weekly)

• Popularity/disagreement is visualized in the software

>> the above two could clash: or somehow make the software configure web2print files such that the two are combined?

• Proliferation of provisional editions

• POSSIBLE: designers make the publication look nice

I think there is some confusion in my mind between transmitting others' words, and this idea of curation. Or maybe not. I'm conflating 'curation' with 'transmitting' maybe. You know, holding a banner for another: both transmission and curation; you decide what you agree with and then you publish it. And it's this process of curation that interests me. When curation isn't simply a personal choice but done out of solidarity. The danger with this project is that this idea gets lost, and it becomes all about debating certain controversial texts in their own right. I somehow have to create an imperative for solidarity (or locate a pre-existing one). I need to find a publication pronto in which to intervene.

And where does curation relate to the 'creation of hybrid documents' that I presented in my diagram? I guess these hybrid documents are simply the edited spreads which will be printed, minus the unpopular text.

The overriding aim is to highlight the dischord that underlies consent. Basically it's Open Sauce, with a more sophisticated way of: - indicating agreement/disagreement - visualising agreement/disagreement

Which is where version control is relevant. Questions to bear in mind while learning Git: - How are documents circulated socially using Git? - How are documents approved of or disapproved of/discarded?

I think my knowledge is still too patchy to create a whole VCS and web to print tool myself - shame, as I wanted a customised one. But I don't just want this to be another Open Sauce, ie, 'artist uses a wiki' -- 'artist uses Git'. I want to draw attention to the social relationships inherent in these tools? No. Actually I want to use these tools to visualise social relationships. (DANGER - being naive about how these tools 'perform' or produce certain behaviour in themselves.)

Goal: make a VCS/publishing system implementing a "beyond yes or no" idea of consent. Editing or version control features to be taken from manifestos on consent - ie, yes, no maybe game.

LEARN what exercises/games are feminists proposing to practice consensual negotiation? Could this be a 'learning game'?

Idea: could the editing/curation process be a more intimate thing, where people are paired with another person and create the document together?

Next steps: Learning about Git to understand how it encodes social relations of approval/discarding of documents; what processes of curation does it take for granted and how could an experimental VCS encode or encourage different relationships?