July 09, 2004

1st July 1661

Today's entry is fortunately quite short and so leaves me some scope to describe my initial approach to doing this topic map and to be a bit more verbose about the mapping.

I propose that the main model for this topic map will be to describe each day as a series of events and at the same time to add descriptive information for participants, places and objects described in relation to each event. For this reason, I have decided to reify each event as a topic.

I will put the topics that represent people and places into separate, parallel topic maps. Fortunately Phil Gyford has already extracted some of the named entities in diary entries and created separate pages for them, so I intend to use those pages as subject indicators where possible.

Topic Mapping This Entry

For each entry I topic map, I will go through the text of the entry primarily in the order in which the text appears. I will use block-quotes to pick out the text being modelled in each section.

This morning I went up and down into the city, to buy several things, as I have lately done, for my house.

This is modelled as a topic of type "visting-event":

[event-16610701-01 : visiting-event 
  = "Samuel visits the city to purchase items for the house. (1st July 1661)";
    "16610701-01"]

I will number the events occurring on a given date in the order of their mention in the diary. For the most part, I expect that this should follow the chronological order of occurrence of the events too. The sort name "16610701-01" is used as a device to get events to sort chronologically in topic map browser applications.

The date of occurrence of events I propose to model as an association, which therefore requires one topic for each day recorded in the diary. So:

[today : date = "Monday 1st July 1661";"16610701"
  @"http://www.techquila.com/psi/date-time/?gDateTime=1661-06-01"]
occurs(event-16610701-01 : event, today : on)

Don't expect that subject indicator URI to resolve to anything (yet) its just a place-holder for now. The gDateTime parameter indicates that the parameter value is a Gregorian calendar date and optional time value.

Finally for this event, I'll model participation in the event as a 3-ary association between the event, the participant (Samuel Pepys) and the place visited. Samuel is none to specific about where he visited, so I'm going to create and use a topic for "London". In general, when I reference topics describing people and places from a diary entry topic map I'll include only the subject indicator for the topic in the diary entry topic map itself and use a separate topic map to keep track of other information related to that subject. For Samuel Pepys, I'm going to use his entry on as a subject indicator. A subject indicator for "London" is a little more troubling because a reference to, for example London on Wikipedia is more focussed on modern London - quite a different place from London in the time of Pepys. Any suggestions for a suitable subject indicator would be gratefully received. For now, I shall create the topic with no subject indicator.

[samuel-pepys @"http://en.wikipedia.org/wiki/Samuel_Pepys"]
[london = "London, England"]
participation(event-16610701-01 : event, 
              samuel-pepys : visitor,
              london : place)

Among other things, a fair chest of drawers for my own chamber, and an Indian gown for myself. The first cost me 33s., the other 34s.

Well, there is no telling what the "other things" are, but Samuel was proud enough of his furniture and fashion to mention them! Rather than combine this information into our "visiting" event, I'll create a separate "purchasing" event to record this. As it is unlikely that these two items were bought in the same transaction, I'm going to model two separate "purchasing" events. As these purchases are part of Samuel's visit to the city, I shall make a suitable association between each purchasing event and the visiting event.

[event-16610701-02 : purchasing-event = "Samuel buys a chest of drawers (1st July 1661)"; "16610701-02"]
occurs(event-16610701-02 : event, event-16610701-01 : during)
participation(samuel-pepys : purchaser, chest-of-drawers : purchased)
[chest-of-drawers : item = "A 'fair chest of drawers'"]
{chest-of-drawers, purchase-price, [[ 33s. ]]}
[event-16610701-03 : purchasing-event = "Samuel buys an indian gown (1st July 1661)"; "16610701-03"]
occurs(event-16610701-03 : event, event-16610701-01 : during)
participation(samuel-pepys : purchaser, indian-gown : purchased)
[indian-gown : item = "An Indian gown"]
{indian-gown, purchase-price, [[ 32s. ]]}

Home and dined there, and Theodore Goodgroome, my singing master, with me, and then to our singing.

Dining is a huge part of the diary - and quite often done in company which makes meals an interesting thing to track. Again, I'll use a different event type to distinguish "dining" events. Simlarly the lesson I shall model as an event, with an additional association describing the "subject" of the event (I'm assuming that I could use the same pattern later to describe events such as sermons or lectures).

The PepysDiary.com site provides useful URI identifiers for Samuel's home in Seething Lane and for his singing master, Theodore Goodgroome.

[event-16610701-04 : dining-event = "Samuel dines with Theodore Goodgroome (1st July 1661)"; "16610701-04"]
occurs( event-16610701-04 : event, today : on)
participation( event-16610701-04 : event,
    samuel-pepys : host, 
    theodore-goodgroome : guest, 
    seething-lane-house:place)
[seething-lane-house : place = "Samuel Pepy's Home in Seething Lane"
  @"http://www.pepysdiary.com/p/1023.php"]
[theodore-goodgroome = "Theodore Goodgroome"
  @"http://www.pepysdiary.com/p/2746.php"]
[event-16610701-05 : lesson = "Samuel takes a singing lesson. (1st July 1661)"; "16610701-05"]
occurs( event-16610701-05 : event, today : on)
participation( event-16610701-05 : event,
    samuel-pepys : pupil, 
    theodore-goodgroome : master,
    seething-lane-house : place)
event-subject(event-16610701-05 : event, singing : subject)

Read this diary entry on pepysdiary.com

Download base ontology topic map (LTM)

Download "people" topic map (LTM)

Download the topic map for this entry.

Posted by Kal at July 9, 2004 10:00 PM
Comments

This is very cool, Kal. I'm going to enjoy following your progress. I really like the basic way you are approaching this, but there are some details I'd like to discuss, some nits I look forward to picking, and some exceptions I will enjoy taking... :-)

Doesn't make sense to do them all now, so let's start with one nit and one exception:

NIT: You've decided to "reify" events as topics and yet people and places, you say, are "represented" by topics. The TMDM (Topic Maps Data Model) makes a subtle but important distinction between these two terms. If we follow the (proposed) standard (and I would like to suggest that we do :-), "represent" is the correct term -- also for events. The term "reify" is reserved for the act of creating topics to represent _other_topic_map_constructs_ (such as associations).

EXCEPTION: You have decided to use Phil Gyford's pages as subject indicators wherever possible. I can understand why (because they have an aura of authoritativeness), but this does go against the OASIS recommendations since those pages were not explicitly intended by the publisher (PG) to be used as subject indicators. That means there is no commitment on the publisher's side to maintaining stable URIs and this devalues them as identifiers. How about approaching Phil to make such a commitment? Failing that I would prefer to see Techquila PSIs used instead (with PG's pages as occurrences rather than subject indicators).

Posted by: Steve Pepper at July 13, 2004 05:34 PM

Re: Reify/Reification - that makes sense to me. The term "represents" seems far more natural than reify anyway.

Re: PSIs - I hadn't intended that these subject identifiers be treated as *published* subject identifiers. However, I will drop Phil a line and see what he thinks about this. That said, it would be good to have PSIs for some historical people and places, and seeing as Restoration London is a favourite period/place of mine I would be happy to commit to hosting such PSIs on techquila...

Posted by: Kal at July 13, 2004 05:56 PM

Good to hear that you will be contacting Phil. Ideally you should try to elicit more than a commitment to keep the URLs stable: Each page could state that it is (also) intended to be used as a published subject indicator in accordance with the OASIS recommendations (available at http://www.oasis-open.org/committees/download.php/3050/pubsubj-pt1-1.02-cs.pdf). Each page should also state explicitly what the published subject *identifier* (i.e. the URI) is for that subject.

Posted by: Steve Pepper at July 14, 2004 09:32 AM