Ticket Review 16 March 2007
Date: 2007-03-23 11:01:19
Agenda: http://icon.stoa.org/trac/pleiades/wiki/TicketReviewAgenda2007-03-23
Chairs
- Tom Elliott (paregorios)
- Sean Gillies (sgillies)
Participants
- Gabriel Bodard (gbodard)
- Hugh Cayless (hcayless)
- Tom Elliott (paregorios)
- Sean Gillies (sgillies)
- Chris Lilley (nantonos, chrislilley)
- Brian Turner (bdturner)
Topics
- open action items
- ticket review for milestone:"Sensible Help"
Issues
- Issue: bug #93
- Issue: bug #199
- Issue: bug #217
- Issue: bug #218
- Issue: bug #220
- Issue: bug #229
- Issue: bug #232
- Issue: bug #238
- Issue: bug #239
- Issue: bug #269
- Issue: bug #277
- Issue: bug #282
- Issue: bug #284
- Issue: bug #286
- Issue: bug #287
- Issue: bug #289
- Issue: bug #287
- Issue: bug #290
- Issue: bug #291
- Issue: bug #292
- Issue: bug #288
Action Items
- sgillies to write a GenericSetup story in the wiki [continuing]
- paregorios to consult community about place types [continuing]
- paregorios to open ticket for an alternative name id and duplicate handling solution [new]
- update 23 March: done; see new bug #294
- paregorios to write CommunityAndCanonicalHelpStory? and req comments via community list [new]
- paregorios to discuss bug #239 with community [new]
- update 23 March: in progress; email sent
- paregorios to engage community in describing (and translating) vocabulary terms [new]
- update 26 March: sent note to pleiades-community list about name-type vocab
Transcript
Please note: this transcript has been edited for clarity.
Introduction
- paregorios: welcome to ticket review!
- paregorios: my fiendish plan is to follow the agenda
- paregorios: so first review of action items from last review, then sgillies will lead us on a spring romp through the tickets for milestone:"Sensible Help"
- paregorios: any objections?
Open Action Items
- paregorios: may I direct attention momentarily to http://icon.stoa.org/trac/pleiades/wiki/TicketReview16March2007#ActionItems
- paregorios: all but two of these are done to my knowledge
- paregorios: sgillies: what's the status on our GenericSetup story
- paregorios: or is the wiki page GenericSetup the answer?
- sgillies: i'm still wrestling with it. it's a long one and i need to email the plone product developers list with some questions first
- paregorios: ok
- paregorios: ACTION: sgillies to write a GenericSetup story in the wiki [continuing]
- sgillies: Jon Stahl has written a product developers doc, and i would like to see about adding some generic setup guidance to it
- paregorios: I think that's a really good idea
- paregorios: the Plone community is trying to improve it's plugin/3rd-party development story
- paregorios: trying to lay in better support, models, etc.
- paregorios: if we can contribute to that effort, I think it would be good
- sgillies: on to tickets?
- paregorios: just a sec
- paregorios: ACTION: paregorios to consult community about place types [continuing]
- paregorios: I want to wait to do this until we've got a bit further on user stories
- paregorios: so leaving it open
- paregorios: ok, now on to tickets
Ticket Review For milestone:"Sensible Help"
Issue bug #93
- paregorios: ISSUE: bug #93
- paregorios: I guess bug #93 is mine
- paregorios: at the core of my work for the coming week
- paregorios: I want to refactor so we aren't using urls and locations to effect the linkage between the context and the help
- paregorios: I plan to use the plone keywords mechanism instead
- paregorios: questions or concerns?
- sgillies: i'm going to presume that everyone is stepped through tickets with us using the "Next Ticket ->" link at the upper right of any ticket page
- sgillies: none
Issue bug #199
- paregorios: ISSUE: bug #199
- paregorios: bug #199 is a specific help document
- paregorios: and is really subordinate to bug #217 (which is next)
Issue bug #217
- paregorios: ISSUE: bug #217
- paregorios: any comments on those?
- sgillies: let's add the dependency
- paregorios: what's our syntax again?
- sgillies: depends on: #n
- paregorios: the dependency for bug #199 is really on bug #93
- paregorios: can't do 199 without 93 being in place
- paregorios: bug #217 is dependent on bug #93 simply because of a part-of relationship
- paregorios: I'm putting them in
- sgillies: lat, lon is specified by EPSG
- paregorios: can you add that comment?
- sgillies: for EPSG:4326
- sgillies: it's frustrating, but there it is
- paregorios: dependencies are in
Issue bug #218
- paregorios: ISSUE: bug #218
- paregorios: as I understand it, we still need landing (index) pages for both "locations" and "names"
- paregorios: sgillies: what was the "issue" with locations?
- sgillies: hang on, we had a collision, and the web ate my comment
- paregorios: greedy beast, that web
- paregorios: sorry
- paregorios thought he was clear
- sgillies: ok, bug #218
- paregorios: the blocker to adding a landing page to locations was the local content model, yes?
- sgillies: let me take this one from you since i'm working very close to it
- sgillies: yes, that, and also because i don't want to lead users into the location container
- paregorios: agreed
- paregorios: I was going to ask
- paregorios: what constitutes good landing pages for locations and names
- sgillies: there's little sense to be made of them in any non-spatial context
- paregorios wants to hide the "duplicate names" solution from users
- paregorios: b/c it will have to change
- paregorios: and anyway, they don't care
- nantonos: why will it change?
- nantonos: isn't there a need for multiple alternate names
- paregorios: right now, names are id'd by the ascii transliteration of the name string
- paregorios: these ids must be unique in the container
- sgillies: primary keys
- nantonos: ah, ok
- paregorios: the solution that worked for this dataset was to dump dupes in the duplicate names subfolder
- paregorios: but as we add more data we will get trips, quads, and more
- paregorios: so the current mechanism will fail
- paregorios: the associations with place and location are handled by references
- paregorios: this is just about buckets for storing the stuff
- paregorios: sgillies: are you therefore taking the landing issue for both names and locations?
- sgillies: paregorios: users will eventually resolve all the duplicate names, right?
- sgillies: ideally
- sgillies: resolve the duplication issue by renaming or deleting them
- paregorios: no
- paregorios: it's simple human practice to use the same sequence of characters for more than one place
- paregorios: I call where I live Huntsville
- paregorios: so do some folks in Texas call theirs
- paregorios: two distinct places
- paregorios: two distinct (but identically spelled) names
- paregorios: nothing to be "resolved"
- sgillies: now i'm with you
- paregorios: we must either adopt something other than name.title for our ids
- paregorios: or come up with a different bucketizing mechanism
- sgillies: the former
- paregorios: +1
- paregorios: whatever we do potentially has an impact on simple and sane urls
- sgillies: we've fallen off of bug #218
- nantonos: I briefly wondered about the wikipedia solution of short page names and disambiguation pages
- nantonos: but perhaps not
- sgillies: second thought, you keep bug #218 and learn about generic setup through it
- paregorios: sgillies: that's fine by me
- paregorios: but I don't think we're off topic for that bug
- paregorios: we may need *another* bug #for a dupes solution though
- sgillies: definitely
- paregorios: ACTION: paregorios to open ticket for an alternative name id and duplicate handling solution
- sgillies: off the top of my head: we keep the existing names, and re-implement plone's id handler to suggest alternatives in the case of already existing ids
- paregorios: nantonos: I think there's probably an issue for users there ... they will eventually want help disambiguating duplicates in, e.g., search results
- paregorios: sgillies: that was what I was thinking
- sgillies: ... that would lead to "xanthos", "xanthos-foo", etc
- nantonos: seems like you can either make all names long, or you can pick a most important name and make the others long, or you can use one level of redirection as needed
- paregorios: picking "most important" would be very difficult
- sgillies: community norms would give us sensible ids
- nantonos: agreed
- sgillies: should give us :)
- paregorios: sgillies: elaborate?
- sgillies: we don't want people entering "xanthos-misc" for the id of the second Xanthos, but it would be dumb to code something that can be managed by community norms
- paregorios: how do we determine community norms for the legacy batlas data
- paregorios: which we are loading piecemeal into pleiades in arbitrary order
- paregorios: but for which we would like to provide persistent urls as soon as practical (as we've already done at the place level)
- sgillies: i suspect that there will be sensible identifiers in the xml which we can use to disambiguate
- paregorios is not sure having users enter ids for names is something they will want to do
- sgillies: let's table this and get back to it after the tickets
- paregorios: right - we have an action item for it
- bdturner: I have some thoughts for later
- paregorios: bdturner: cool
Issue bug #220
- paregorios: ISSUE: bug #220
- paregorios: this is actually at the top of the dependency tree we trundled into with bug #93
- nantonos: Why is "change the linking mechanism from url fragments to keywords or relations" better?
- paregorios: the way I'm doing it now the url fragments (unique ids in containers) in the path have to match at a certain point
- paregorios: with
- paregorios: class and attribute names in the content object models
- paregorios: ugly and cumbersome
- paregorios: and a single help topic can't then be reused in multiple contexts
- nantonos: okay
- sgillies: http://pleiades.stoa.org/help
- paregorios: I've added a dependency relationship with bug #93
- sgillies: cool
- paregorios: anything else on this bug?
- sgillies: would be nice if it aligned with our new vocab views
- paregorios: can you add a comment to that effect on bug #220?
- sgillies: so that a field's help is at http://pleiades.stoa.org/help/class#field
- paregorios: for helpish content that maps well onto a list of terms, I think that's a good idea (i.e., reference-style help)
- paregorios: for concepts and tasks, we'll need something else
- sgillies: agreed
- paregorios: ready to proceed to next bug?
Issue bug #229
- sgillies: ISSUE: bug #229
- paregorios: sgillies: I think I asked you before, but don't remember the answer
- paregorios: what's needed here?
- sgillies: a generic setup config file
- sgillies: ask me when you get to this one
- paregorios: ok
- paregorios: that's mine
- sgillies: trivial, but will need testing
Issue bug #232
- paregorios: ISSUE: bug #232
- paregorios: what do people think of this idea?
- sgillies: +1
- nantonos: would it imply a need for help aggregation?
- sgillies: the best could be distilled into "official" portal help
- bdturner: does this mean I'll have a list of personal helpfiles I used often?
- paregorios: it would mean that you'd have a collection of helpfiles that you authored
- bdturner: oh!
- paregorios: they could either be drafts ... in which case they'd stay your personal stuff
- paregorios: or they'd get workflowed to published status (by someone with appropriate role permissions), and then they'd be global
- bdturner: useful...but I am concerned about "clutter"
- paregorios: they'd need to be marshalled into some kind of central view
- paregorios: there are two models with plone, aren't there
- bdturner: clutter = to many files and folders and buttons etc
- bdturner: OT I like it
- paregorios: bdturner: I think we'd try to have a pretty clean view of a list of documents, with some kind of sensible top-level organization
- paregorios: but we'd also put fairly discrete links, or a single portlet with links in it, on other pages ... providing access to only those help docs (wherever they are) that are relevant to the user's current work context
- bdturner: implied... I just wanted to make the point... especially for people (like me) not familiar with plone... to many options can be skeery
- paregorios: yes, I know the front page is still an issue there too (I think we have a ticket for that)
- bdturner: probably off topic now
- bdturner: re: front page - yes
- bdturner: definitely off topic now
- paregorios thinks the default plone layout and portlets collection is mostly about showing off what plone can do
- bdturner: makes sense
- chrislilley recovers from blue screen :(
- sgillies: yeah, we'll do away with some of the clutter soon
- paregorios: chrislilley: where did you check out
- sgillies: bdturner: i think the calendar is not very useful
- chrislilley: I was only gone for a couple of minutes
- paregorios: you were right about needing help aggregation, which I think we can pretty easily do on the basis of plone keywords
- chrislilley: yes
- paregorios: using Plone "smart folders" and workflow states
- bdturner: calendar... but I'm mostly concerned with the redundant buttons and multiple folders which I think can be collapsed into one or two at most....again probably tangential
- chrislilley: if multiple people create help then a help query has multiple sources to draw on
- paregorios: the navigation portlet is problematic
- chrislilley: thus, how is it aggregated
- chrislilley: or, you can say there is system help and user help
- chrislilley: just two sources
- paregorios: chrislilley: you mean the collisions problem again?
- bdturner: re: multiple help files... redundancy? will show two ways to do the same thing?
- chrislilley: and i don't benefit from your help until it gets promoted to system
- bdturner: sorry... redundancy
- chrislilley: yes
- bdturner: re: two ways of completing same task... useful... confusing to new users???
- paregorios: one model for solving this is to have "canonical" help and "community-generated" help
- chrislilley: in other words do people only see their own local help, so it's like a prototyping service, or do we see everyones?
- bdturner likes idea but is playing devil's advocate
- bdturner: ditto Chris' comment
- paregorios: we *could* (depending on cycles) treat as follows
- paregorios: help gets surfaced in one of two ways
- paregorios: field-level reference help (that is marked as "published" and "canonical") gets links generated to it next to widgets in appropriate views
- paregorios: higher-level, but context-relevant help of any type shows up in a portlet
- paregorios: such help if marked as "canonical" would be at top in portlet
- paregorios: then under a subheading, you'd get anything that was at the community level
- paregorios: part of the process of moving something from community to canonical would be disambiguation, consolidation and relocation to a central place
- paregorios: therefore a function of the managing editors, development team and others press-ganged into helping
- chrislilley: that sounds like a workable plan
- bdturner: agreed
- paregorios: how about I write this up as a story in the wiki and hurl it at the community list for reaction?
- bdturner: +1
- sgillies: sounds good. shall we go on to bug #238?
- paregorios: it won't slow work on the other help-related tix as alot of this is business rules and query criteria
- paregorios: ACTION: paregorios to write CommunityAndCanonicalHelpStory? and req comments via community list
Issue bug #238
- paregorios: ISSUE: bug #238
- paregorios: I don't think this is hard, once I figure out how to force execution of a field validator
- bdturner: re: 238 ... same code as when I try to save something and use a "." in the file name??
- bdturner: computer says NO
- paregorios: there are plone hooks for this, I just haven't done it before
- paregorios: bdturner: same general idea
- sgillies: yep
- sgillies: we could even spell it out as a regex
- paregorios: +1
- sgillies: for those that like regex :)
- paregorios: it's what I'll use in the code
- sgillies: i assume digital humanists are familiar with regular expressions
- paregorios: it's pretty easy, it has to be [a-zA-Z]+
- paregorios: if they're not, they need to become so
- chrislilley: um, hang on there
- paregorios: sorry, no that's for nameTransliterated
- chrislilley: all names are restricted to ascii?
- paregorios ducks chrislilley's cudgel
- chrislilley waves the Unicode stick
- paregorios: chrislilley: no, I mistyped
- paregorios: for bug #238 I must *exclude* various types of punctuation
- paregorios: leaving everything else legal
- chrislilley looks happier
- paregorios would like to move on to bug #239 before he gets in further trouble
Issue bug #239
- sgillies: ISSUE: bug #239
- chrislilley: are people going to want different transliteration schemes?
- hcayless: I think as long as you document what scheme you're using, it'd be fine
- paregorios has been operating under the assumption that there would be only one
- paregorios: per language/script combination
- chrislilley: seems there are two options
- chrislilley: one is to disable editing and autogenerate the transliteration
- chrislilley: the other is to enable editing, but pre-fill with a suggested transliteration
- paregorios: yes
- paregorios: the current data employs a single, uniform transliteration scheme for classical Greek
- paregorios: with pleiades, I would like to shift the focus to the actual orthography in the original languages/scripts
- chrislilley considers waving the etruscan stick
- paregorios asks that the Etruscan issue be tabled for about 5 minutes
- chrislilley: +1 to actual orthography
- bdturner: re: uniform transliteration - presumably into English language? What of a German transliteration ... or worse... Canadian
- paregorios: not into English language
- paregorios: into Roman characters
- sgillies hoists etruscan cinary urn
- bdturner: check
- chrislilley: http://en.wikipedia.org/wiki/Iberian_script
- paregorios: you're thinking of things like the heinous Pauly practice of slapping a German case ending on an otherwise Greek name
- paregorios would call that a "Germanized" name
- paregorios: i.e., *not* the ancient name itself
- paregorios: there are a couple of reasons, in my view, for doing transliterations:
- paregorios: 1. simple sorting
- paregorios: 2. user familiarity (helps users who don't have the original language)
- chrislilley: I'm wondering about the capricious choice of b, p, t for transliterating Iberic, depending on what point someone wants to make with their etymological analysis
- chrislilley: same in latin for u and v of course
- paregorios: the managing editors, with community consultation, will have to undertake the specifications for what can appear in both attestedName and transliteratedName
- paregorios: on a language/script by language/script basis
- paregorios: at present, we only allow
- paregorios: grc, la, grc-Latn and la-Grek
- hcayless: I think you're better off with one scheme where possible, and auto-converting
- chrislilley: "at present" - what do you plan to allow later?
- paregorios: the legacy data sets an existing transliteration practice for those four
- paregorios: what to allow later is also for community discussion and eventual sanction (or not) by the managing editors
- paregorios: I'm about to start a list discussion on WorkFlowStories
- paregorios: I will ask community members to contribute short descriptions of example additions/modifications/corrections they'd like to make to Pleiades data
- paregorios: I hope that those who are carrying a torch for other ancient language/script combinations will surface the issue in that context
- paregorios: we'll then prioritize and categorize the stories, generate tickets and milestones and move out
- paregorios: hcayless: that has been my assumption
- chrislilley shelves his torch for now
- bdturner: count me into auto-converting
- bdturner likes it when things are done for him
- paregorios: chrislilley: for languages closely related in space, time, cultural or linguistic terms to Greek and Latin, I don't think
- paregorios: there will be much trouble adding support quickly
- paregorios: so long as they have a complete unicode subrange and there is a recognized transliteration system
- paregorios: adding support for those that don't meet that set of criteria will require more study
- paregorios: and coding
- paregorios: getting a mechanism in place for automating transliteration (the purpose of this bug) is a prerequisite IMO
- paregorios: I intend to implement the mechanism in a way that is language/script aware (on the basis of user input) and readily extensible
- chrislilley points to http://linguistics.berkeley.edu/sei/USR.html#n37 and http://www.omniglot.com/writing/iberian.htm then leaves it be
- paregorios: thanks
- chrislilley: did the question of editable transliterations get resolved?
- paregorios: hcayless and I think that they should not be editable
- paregorios: bdturner too, no?
- bdturner: yes...
- chrislilley: hmm
- chrislilley: I foresee complaints when its not the transliteration the user expects
- bdturner: but I recognize Chris' problem ... especially if we intend to be worldwide
- chrislilley: but it will give you more consistency for the common cases
- hcayless: I can imagine that for some scripts, editability might need to be turned on...
- paregorios: chrislilley: would a "report erroneous transliteration" bug/link help?
- chrislilley: yes
- chrislilley: thanks
- paregorios: hcayless: I suspect you're right
- hcayless: but for Greek at least I'd argue for picking one
- paregorios: there may be some lang/script combos for which a deterministic conversion from the written to a standard/scholarly transliteration (or representation) scheme
- bdturner: so... normally we will have autoconvert... but with the option of a user to override the automated and add their own transliteration?
- chrislilley: that sounds good
- paregorios: may be problematic ... a choice between options may be predicated upon an "understanding" of what's meant by characters
- bdturner: well code in a block so they can't add the @ sign
- sgillies is still here, listening quietly
- paregorios fears that there will be pedants who refuse to comply with common practice
- hcayless: I would not make override an option
- bdturner: oh ... there will be... no matter what we do
- paregorios: and will therefore insist on modifying a perfectly compliant transliteration to fit their idiosyncratic ideals
- hcayless: unless there are scripts which have problems being auto-transliterated
- hcayless: and only in those cases
- paregorios: hcayless: so the option is only surfaced for lang/scripts that site policy wants it surfaced for?
- hcayless: that's what I'd suggest
- paregorios: otherwise, you have to content yourself with a "report error" link that allows you to lambaste the managing editors in extended prose
- bdturner: but how will we know what those are?
- paregorios: two ways:
- paregorios: 1. the managing editors say so (or not) on the basis of their best judgment and consultation
- paregorios: 2. the managing editors reconsider an earlier decision on the basis of the amount of hate mail they're getting
- paregorios: (i.e., the error rate reported by the community)
- paregorios: maybe the link is: flag transliteration as bad (and you have the option of attaching a raging comment if you want)
- bdturner: or the managing editor ignores the complaint and gets a new email :)
- paregorios: how we decide to go on this is pregnant with implications for future decisions of the same class
- bdturner: yeah... we're basically deciding what we'll allow the user to do
- bdturner: Lincoln: can please some of the people.......??
- paregorios: my concern in this context is to minimize unnecessary and potentially workflow-clogging and search-frustrating churn while still addressing needed flexibility
- sgillies: paregorios: is a wider discussion needed to move on the ticket?
- paregorios: this probably needs airing at the community level
- sgillies looks ahead to 269
- paregorios: bug #239 does not block anything
- paregorios: ACTION: paregorios to discuss bug #239 with community
- paregorios: ok?
- sgillies: cool
- bdturner: yes....
- paregorios: chrislilley?
- chrislilley: ok
- bdturner: absolutely
Issue bug #269
- paregorios: ISSUE: bug #269
- sgillies: done :)
- paregorios: :)
- paregorios: mark it so
- paregorios: and lead us onward
Issue bug #277
- sgillies: ISSUE: bug #277
- sgillies: i've accepted
Issue bug #282
- sgillies: ISSUE: bug #282
- paregorios: I've accepted bug #282
- sgillies: i wonder if Membrane or other new membership products provide this?
- sgillies: worth a look
- paregorios: I think there's a standard way to put some content in the portal that gets shown there
- paregorios: it'll be the same for all members (probably translatable with LinguaPlone in place, but that's down the road)
- paregorios: but I'll take a quick look so we know
- sgillies: yes
Issue bug #284
- sgillies: ISSUE: bug #284
- sgillies: meaty
- paregorios: point is: should we maintain atlas practice of subaudibilizing (at the top label-level) ethnic names when there are attested geographic names
- paregorios: saying "Alabama" for our title for the place, not "Alabama/Alabamian"
- sgillies: i agree
- paregorios: only using the ethnic for a place title when there is no geographic
- paregorios: I think modifying the code is pretty easy ... I looked at it before I wrote the ticket
- bdturner: +1
- paregorios: so I"m still happy to accept this one too
- paregorios: unless you think I'm too heavily burdened ;)
- sgillies: careful management of placeful associations should prevent this from happening
- paregorios: explain
- bdturner: though ethnicities without geographic references should be in a separate category methinks
- paregorios: those are just names
- chrislilley: and they are rarely points
- sgillies: yes, what chrislilley said
- sgillies: ethnic names will be associated with regions (when we add them to our system)
- paregorios: the locations associated with ethnic names are the (polygonal) extents of the range or habitation or territory of the people known by that name
- chrislilley: I can think of plenty of maps with ethnicities strewn around to roughly indicate areas
- chrislilley: eg 'germanic tribes' maps
- paregorios: BAtlas has them by the legion
- paregorios: I don't think the above discussion of ethnicities precludes implementing the proposed solution for bug #284
- sgillies: i agree with you
- chrislilley: i wasn't disagreeing
- paregorios: understood
Issue bug #286
- paregorios: I've accepted bug #284
- paregorios: this won't get done this milestone, but ...
- sgillies: defer 284 then?
- paregorios: let's revisit next friday
Issue bug #287
- sgillies: ISSUE: bug #287
- sgillies: looks good
Issue bug #289
- sgillies: ISSUE: bug #289
- sgillies: i'll accept, and we'll try this out on the pleiades-dev site for a while
- paregorios: ACTION: paregorios to engage community in describing (and translating) vocabulary terms
- paregorios: +1
Issue bug #287
- paregorios: ISSUE: bug #287
- paregorios: I've accepted
Issue bug #290
- sgillies: ISSUE: bug #290
- sgillies: same there
- sgillies: experiment with performance on the dev site
- paregorios: +1
Issue bug #291
- paregorios: ISSUE: bug #291
- sgillies: agreed. we're now indexing the vocab key and captions
- sgillies: accepted
- paregorios: right and users no longer see the keys in the view s
- paregorios: so they don't need to see them in the help
Issue bug #292
- paregorios: ISSUE: bug #292
- sgillies: accepted
Issue bug #288
- paregorios: ISSUE: bug #288
- paregorios: that's a css thing
- paregorios: accepted
- sgillies: i'll look into that
- paregorios: oh ok
- chrislilley: beware relative positioning
- sgillies: that's weird
- paregorios: I suspect it's classes that are (mis/re)used
- paregorios: inherited from plone
- paregorios: and something I did in trying to play with our border colors in the customized bits above
- paregorios: do some poking around with firebug
- paregorios: thanks for all the great feedback!
- paregorios: this was very productive
