Peter Ellis » Lecture Notes

Notes: Electronic Piers Plowman: Implementing an Edition of a Six-Hundred-Year-Old-Poem for Twenty-First Century Students

Peter Ellis — Fri, 07 Mar 2008 23:56:42 +0000

These are my own notes from the Research Conversation about representing Piers Plowman today, March 7. Presented by Terry Brooks and Miceal Vaughan. (Note (3/11/2008): I went back in and cleaned some of the formatting up on this, since apparently Windows Live Writer is not quite as consistent as I’d like.)

Miceal:
- Work is an electronic edition of the poem aimed towards beginning students
- 14th century poem in 3 distinct versions – 50+ manuscripts!
- First version ~1360s – author is uncertain, could be multiple issues
- Trying to build an archive of all surviving manuscripts of the poem
- Want to produce the poem so that it’s accessible in different formats (wording/spelling formats)
- Review of different versions of first four lines – spelling differing from Middle English to modern with different representations of letters: þ (equivalent to “th”), for instance
- Miceal has been working on this for roughly 15 years.
- Wants to present clean, uncluttered version so that it can be worked with directly
- Had HTML version created for students to work with created by people knowledgeable in HTML
Terry:
- Problem: Content of the poem is being held hostage by a fixed presentation format. Wants to divorce the format from the content.
- Given the original HTML version; first had to go into the source and create a parser that parses out content. Has to also deal with inline commenting. Wants to “unfreeze” the content and put it into a general format.
- XML structure: Section (prologue, passage) made up of many lines. Break the poem down to a word-by-word unit level, which could have separate spellings and spacing requirements and (of course) the value of the word itself. Also has word-by-word footnotes, which have author and line information.
- Problem with representation in XML: HTML escape characters can’t cooperate well with XML parsers. Solved by HTML post-processing that replaces placeholders with the proper characters.
- This was initially done so that it would be fixed, but editing was a need – created an XML editor to allow for these attributes to be changed without knowing the XML itself.
Current XML version focuses on the Knott-Fowler and Vaughan versions of the poem, and encodes line-by-line metadata on each version.
How does one flag the reader that there’s content hidden behind a particular word? Their tactic: change the icon.
How does an expert in Middle English want to read more than one text? Side by side, one above and one below? Actually, it turns out that they want two versions of each line presented together rather than side by side.
Question to audience: How would you build an application so that you can compare two versions (from an information architecture standpoint)?
Question from audience: How do you make this project expandable beyond Piers Plowman?
- Some of these strategies may be reusable.
- Micael: Trying to give access to a variety of versions.
- Terry: Yes, aiming at a generalized information architecture
Question from audience: Why not database-driven instead of XML?
- Terry: Portability.
Question from audience: How do scholars process and comment upon these works? Does literature exist that researches how this is done?
- Micael: No – everyone comments on these works differently.
- Terry: The XML editor spurred the same question – how do people edit these things? Feedback regarding how Micael edits led to editor program changes.
- Micael: There may be a possibility of redefining how commenting is done via the editor.
Question: Pedagogically, is this the same as you would have had students do by looking at two versions side by side?
- Micael: They would have gotten one version. Wants this to be something usable by anyone interested, but also for people to be able to understand the significance of different versions of the same work.
- Terry: I would ask the same question, different angle: How do people read these poems?
- UI design: Micael doesn’t want underlining or coloring of words – wants the data behind the document to be transparent
Micael envisions this project as an archive.
Audience: Interesting part of this is learning from the juxtaposition of different texts. But you’ve got a grid of future/current technologies and activities to consider.
Micael: Johns Hopkins not interested in keeping the electronic version of this information, but others might be interested in coordinating this.
Audience: This would embrace different learning styles that look at the same work from different angles. Example: Translating works, you might work an existing translation to help your translation efforts to make sure you’re being accurate and to help you understand it. Someone else might benefit more from struggling through the translation process without any checks against existing translation work.
Audience: This XML structure really doesn’t seem that portable, since it’s dependent on whether other documents use the same attributes.
Terry: Correct – not sure how generalizable the schema is.
Audience: This would have huge problems going across languages (French, Spanish, etc.)
Micael: This representation scheme would work for words, but not for prose.

Notes: Using Uncensored Communication Channels to Divert Spam Traffic, January 31, 2008

Peter Ellis — Fri, 01 Feb 2008 00:58:56 +0000

This was a presentation given by Benjamin Chiao from the University of Michigan – he’s currently a PhD student at their Information School, but also has an economic background, which is where much of this talk was couched.

What’s the point of solving spam problem? Less time sorting spam, less economic cost for blocking spam, customers spend less money
$10 billion/year spent on spam related technologies
What is uncensored/open channel? keep inbox filters, no filters in special folder, guarantee delivery of messages into folder
Properly tagged messages will automatically be assigned to a folder/label
No new technological infrastructure required and fully reversible
Existing mechanisms to prevent spam: legal punishment, filters
Proposal of the open channel: decrease benefits of spamming by decreasing the number of recipients
Economics: micro-economic model shows open channels increase benefits to recipients, advertisers
This is not a unique mechanism – Chiao compared it to TV shopping channels: you don’t have to watch, but the information is constantly there
Open channel is like web sites – anyone can post
Not excluding the possibility of search within the open channel
Sender tags sent messages (as being part of the channel? This wasn’t clear in the talk)
The definition of spam used here specifically targets unsolicited commercial mass e-mails – no other message types are considered here
Current spam volumes are between 80-90% of total network traffic – 40% advertise medications, 19% is adult content, 41% other (according to Evett 2006
Spammers continue because they are economically supported – there’s a point where the supply of spam must meet demand
Why do we need open channel? Why not just search for the content via existing search engines? Sites selling these products disappear too quickly: 30% of domains created die within a day (according to MessageLabs 2005)
Spammers need to keep pushing information to inboxes because they must move rapidly due to legal reasons
60% of spam messages are sent by zombies – computers hijacked for the explicit purpose of sending spam
The CAN-SPAM Act has essentially legalized spamming
The open channel proposal separates the current e-mail ecosystem into two ecosystems – one “open” (the proposal) and one “traditional” (the current model)
Audience observation: this system assumes that EVERY e-mail system implements the open-channel concept
Current technology already partially implements this idea (sort of)
Spammers might be happier on open channel!
This is still a theoretical idea
Essentially create two channels: one open and one censored (I’m not clear on whether the “channels” are analogous to the “ecosystems” mentioned above)
E-mail recipients opt in to the open channel in order to maximize their own utility
The sender gets its current revenue from the advertising charge times the number of mails received
The sender’s current cost is the constant reestablishment of sending channels (zombies)
The open channel attempts to establish equilibrium between advertisers and receivers of spam (note that advertisers, senders, and receivers are independent parties)
There is not just a supply curve but a demand curve for UCM
The open channel method induces UCM to move out of the current e-mail system

I’m not sure Benjamin gave sufficient background to make any of us fully appreciate the idea – there’s two problems with it that I can see: first, it exists within the reality of economics, not the reality that we commonly deal with. Thus, it’s governed by the same economic laws that give me such a headache in PB AF 594, and understanding the concept requires a suspension of our own realities in order to appreciate the laws that govern the proposal. The second problem is that it’s not clear how this can be implemented within the current system. Is this a system that merely adds a tag to all messages that identify it as open-channel or “traditional”? How do you physically separate the two ecosystems without actually modifying the current e-mail structure, and how do you enforce proper usage of both ecosystems? An honor system in which we assume that the senders, the receivers, and the advertisers are all working to maximize their own utility (basically their net happiness) is perfect in economic theory because economic theory establishes that everyone will strive towards some theoretical maximum benefit, but in reality, it just doesn’t seem possible.

There was one thing that I want to follow up on – Benjamin mentioned the Attention-Bond Mechanism (Loder 2006) in his talk, so I’ll have to look up exactly what that entails (it’s a concept related to the acceptance or rejection of e-mail messages).

Notes from Central Debates of Sustainable Design, January 15, 2008

Peter Ellis — Thu, 31 Jan 2008 23:02:17 +0000

Ann Thorpe, author of The Designer’s Atlas of Sustainability, gave a talk entitled “Central Debates of Sustainable Design” as part of the UW’s Luce lecture series. My notes from this are below.

Wanted to cover in the book where people come from when approaching sustainability and how to “do” sustainability
Book systematically and visually presents the concept
- “The making of” – the book started in 2001
- Based on the principle of “If worth doing, it’s worth doing badly” – had to start somewhere for the idea
Central debates: responsibility, pace (of adoption), scale, operation, and appearance. The talk focused on the first three central debates.
Designers rarely have time to get up to speed if they don’t know about sustainability
The market is not the same as the overall economy
Natural resources have different prices
“Let the market decide”
The operational spheres of nonprofits, private, and public organizations all overlap with the economy
It appears cheaper to destroy natural or societal resources than it is to preserve them according to the market
What responsibilities do designers take across a market economy?
Nonprofits will be seen as having a potential for a proactive stance in promoting issues
Part of the problem is how we take things from the ecosystem and then redistribute it
We don’t see the costs for the global distribution of produced materials
Different things work at a different pace – art/fashion, communication, infrastructure, culture, nature (this list is actually sorted fastest to slowest in terms of rate of change)
Commerce is starting to control the pace of change
Much as we want to push change, we need stability in the (economic) system
Change takes three forms: physical, economic, and cultural
Audience question: sustainable costs more – can we make it cost less? Do cases of this happening exist?
Answer: marketplace tools are a solution here. There are some cases where this has happened.
Things are cheap in monetary terms that aren’t in sustainable terms – this is a systemic problem
“Be an active citizen” to make sustainability viable – knowledge is power
Sustainability is complex and depends on context of values
How might open source play into sustainable design?
Audience: We do sustainable buildings, but it “doesn’t look good”

Information Architecture Panel: January 10, 2008

Peter Ellis — Thu, 31 Jan 2008 22:01:54 +0000

The iSchool held a panel of practitioners in the field of Information Architecture as part of iCareer Week at the beginning of the quarter. These are my notes from that panel (interpreted without quite as much of the fresh context in my head as I usually like, I must admit).

Recommended courses: 530/540 (taxonomy/technology); project management; research methods and analytical skills; coollaboration/teamwork; 580s; User Centered Design (UW Extension/UW Educational Outreach)
When in class, always ask the question: “What’s the point of this?’ Also ask, “How can I communicate the value of this to someone not familiar with the concept?”
What kind of information architect do you want to be?
“What other value can I add to the degree?” (ask this)
Mike Crandall: What functional subspecialties are there in IA? What is IA?
- metadata, user experience, user research, usability, analytics, wireframes, data modeling, interface/interaction design, evaluation
- Major groupings within IA: taxonomy, HCI, design (visual and interaction)
- There are also “innies” vs. “outies”: internal and external consultancies in IA. The contrast here is one of a mother who takes care of the kids and an ER doc who does triage. This is the contrast between a consultant for a company and a consultant for an agency.
- If you want to get into programming as part of the MSIM degree, go for concepts, not languages
- It’s difficult to be a developer and an information architect.
- It’s important to be able to talk to people who understand how the system is built – hence why programming can be important.
Mike Crandall: What kinds of tools do you use?
- Outlook/Excel/Office, mind mapping software, Illustrator (some), InDesign, workflows
- “We’re consultants first” – need to be able to advertise and deliver IA. How to express that? “Deliverables, wireframes…”
- “Tools were not a big concern [in work] – I had the underpinnings.”
Mike Crandall: Looking for a job – what did you do in your job search? How did you find IA-related jobs?
- Join a professional organization related to IA and stay informed
- Look for a job using your own personal network – go through the people you know
- Make your own projects while you don’t have a job and build it. Create your own portfolio.
- Get into a company that needs what you want to do and do it (note that this may not match your “official” job title!)
- Brush up on resumes and cover letters
Audience question: how do you present stuff that’s not really done?
- Get to the level of “I feel good about this piece of work” and give the context of the assignment
- Depends on the type of IA you want to do
- You’re attempting, in your portfolio, to show how you synthesize a large amount of work
- Mike Crandall: process is the important thing: you sell the process, not the work.
- Don’t be afraid to say you don’t do something because it isn’t your strength!
Mike Crandall: What is your next career step?
- CEO! Getting more of the science behind the ideas, getting more in front of clients, practicing current work, “getting good”, balance design with working with people, create a collaborative process, start their own company, become web director, learn business skills, learn team management, work on motivation (self and others), get more management/oversight experience, work on client/account management
Audience question: What do you hate?
- Being rushed, repetition, wireframes (sort of)
wireframes are breaking down – high level of interaction
Special Interest Group lists are valuable
Audience question: when did you finally feel confident?
- When working with first client
- Focus on the user is your selling point
- Fake it – say stuff with confidence, even if you have no clue what you’re doing
- Know how to figure things out

Research Conversation: Personal Information for A World As We Want It to Be

Peter Ellis — Sat, 17 Nov 2007 00:26:39 +0000

William Jones, one of the professors at the iSchool, gave a really interesting talk about the idea of personal information management and how to improve our ability to find the information we need. Jones is one of the lead researchers for the Keeping Found Things Found project, which is a project that I’ve had some interest in since I discovered it through my research on the iSchool itself.

Some notes from the presentation:

Why do we have folders?
- From the audience: to organize data.
  - Why do we organize data?
    - To find/locate information.
- As a quick reference into the materials we need.
- As content metadata
Search on our own machines gives us the ability to get stuff the same way as on the Web, so why would there be resistance to this?
Audience member observation: There’s a difference between finding things and finding new stuff
Folders are a part of our interaction with data
Why do people use folders in so many diverse ways?
The Web is becoming an extension of ourselves (and of our personal information)
Capturing information is now very easy
Storage is now very cheap
Search makes retrieval of information easy (if it is properly indexes and if there’s some form of version control – search does no good if we’re looking for old versions of things we already have)
Information fragmentation – the idea that our information is now incredibly spread out – is a more recent problem than that of information overload, which has existed, one could argue, for centuries
Keeping Found Things Found project did three major studies:
- How people keep information
- How people re-locate information they have
- How people organize their information
There is a lot of diversity in the way that people organize their information – why is this?
An audience member gave an example of using e-mail instead of favorites or bookmarks to manage their web site. When asked why, they explained that they didn’t want their favorites list to get too long or unmanageable.
What about the recall of information? KFTF participants were given a list of information they had accessed 3+ months ago and asked to relocate it quickly using whatever method they wanted. They were only given five minutes for the task. After that five minutes, it was found that there was a 95% successs rate in finding that information based on a list of particular conditions (what those conditions were wasn’t discussed in the talk). However, there were some issues with people trying to remember where that information was stored. It was also noted that “Do nothing” methods – where people had made no prior note as to where the information was located (methods like Google searching) won out over bookmarks and most other methods of information search and retrieval.
Fourteen participants were asked to give a tour of their folder/information organization on their computers. For every single participant, there was something where they said “this shouldn’t be here”, and a small number even had to stop the demonstration to move the information to the correct location.
An idea Jones suggested was that old information should slowly fade from view – it doesn’t get deleted, it just isn’t visible.
It’s easier to pay the small cost of not being able to find things immediately than to pay the larger cost of having to reorganize or clean out our information resources.
An audience member noted that economics can play a big role in how information is organized, especially in a work environment – if we get paid to do things quickly, our information organizational structure better make things easy to find!

Steve Krug at Adobe Seattle

Peter Ellis — Tue, 30 Oct 2007 16:54:11 +0000

I had the pleasure of hearing Steve Krug, author of the book Don’t Make Me Think: A Common Sense Approach to Web Usability, at the Puget Sound SIGCHI meeting on the 25th. This is a great little book that’s been on my bookshelf for a while. Steve’s a great speaker and a deep thinker about the subject of web usability, so it was an interesting talk. Some of his key points:

There are two things that every designer overlooks: “You are here” indicators and page titles.
“You are here” indicators need to be louder than you think they need to be in order to grab attention. These can be in the form of tabs that are shaded to match the page background (he uses StumbleUpon as an excellent example), or in any other format that makes the indicator “pop”. Steve has a confessed bias towards tabs, though.
There needs to be a top-level “Home” option – simply having the logo link back to the home page isn’t enough. This is so that it’s easily locatable and so that people always know where they are in relation to the main page. If you use subnavigation under category tabs, center the subnavigation under the tab.
Prominent, well-placed page titles are a must. Steve says that “if I look at a page from 50 feet away, I should be able to guess the content of the page”. This doesn’t mean that it has to be the biggest word or even the boldest word on the page. Rather, it means that the page title needs to be well-placed at the top of the content space. It should take advantage of its prominence and its location on the page. He offers up the idea that WYCIWYG (what you click is what you get): in other words, if you click on the link, the page title and the text of the link should convey the same idea. This doesn’t mean that a link named “Contact Us” links to a page with the same title; you could use a variant such as “Get in Touch With Us”, so long as the main idea is conveyed.
“So Steve wants all sites to look the same?” No. There are exceptions to these rules (entertainment sites and sites that are meant to be puzzling, to name a couple).
The best piece of advice I’ve heard in a while: if something on a web page doesn’t work for a group of people using the site, that’s not an indicator that you have to scrap the design and start over. Steve is a big advocate for making the smallest tweak possible that makes the site more usable.