BAV Chigi D.V.71

As mentioned, most of the usual updates are going through Twitter, @cokldb, but Chigi D.V.71  from the Biblioteca Apostolica Vaticana  is far to complex to explain in 280 characters.  First some numbers, this is the 18th MS from the BAV, and it was made in Toulouse, though liturgical Use of Rome, circa 1335-42.   After the January and February calendar pages there are Latin passages, written in red, on calculating the date of Easter.

The calendar is unusually large in several measures.  Each entry takes 2 ruled lines, with the dominical letter, golden number and roman number on the first line and the roman part stretched across both, sometimes even three.  The stretched Roman part is not unusual, but the double-line entry is unique.  This layout means that each month takes 2 full folios, 4 sides, leading to a calendar 24 pages long.  The decoration is rather sparse, but standard mid 14th Century, with a border and a few leaves on the first face of the month and nothing on the other three.  This layout allowed the scribe to load the calendar with saints, there are approximately 480 distinct feasts, with 9 dates carrying three.  August 22 has four distinct feasts: the Octave of the Assumption(469), and three distinct martyrs, Thimothy(478), Hippolytus(471), and Symphorian(477).  For context, a fully populated calendar has between 365 and 380 saints, and the previous MS with the densest calendar hit 401.  Not surprisingly, this calendar has a number of hapax saints, 30 new saints and 48 unidentified ones.

The calendar decorations are less unusual, rod borders with sparse foliage, some of in intruding into the text block.  There is no KL to start the page, in its place there's a small miniature of the labor of the month, backed with burnished gold.  To ornament the dominical "A" there's some small amount of red or blue penwork.

Note that this MS makes approximately 250 devotional calendars added to the DB since December 2015.

(DB Id: 381)

4 Saints Verified!

With the addition of Vatican, we have verified that 4 saints in a single date now loads and processes correctly.  Taking a step backwards, this is a Use of Rome hours with a Calendar from Aquilla (probably) that may have been owned by Queen Christina of Sweden.  The calendar is either unfinished or the scribe was confused, each entry is missing the initial letter, eg. "pyphania do" for Jan-6, missing the initial "E".  For testing, though, August 22 (see f.6v below) is the critical date, having entries for Sts. Hyppolitus (471), Timothy (478) and Symphorian (477), and the Octave of the Assumption (469).  This was entered into the new 4-saint Excel template  and it loaded and displayed cleanly!

(DB Id: 149)

4 Saints!

A bunch of changes have been underway in the background.  First a big procedural one - I'm going to be trying to post shorter pieces about newly added MS to twitter: @cokldb.  I've added about 10 MS recently, and many of those are mentioned there.  When I have a longer discussion, I may still post here.

In the backend, we now support up to 4 different saints per date!  This required changes in a bunch of places, and may not be perfect, but seems to work.  There are not many books in the DB with more than 2 per date, about 30 entries have 3 saints and 3 have 4.  This is one of the places that need work, I scanned the existing entries for ones marked as "three saints" or "four saints" but if I was inconsistant, I may not have found all of them.  The loader/excel template was also updated to handle 4 entries and 4 colors.

The current work, yet undeployed, is to get some statistical graphs into the date detail pages.  Ideally a pie chart of "blank vs populated" and then a breakdown of saints in the populated ones.  The data is in place, what is taking work is a good display

Backlog of Achievements

Having been busy, but distracted, I've neglected to update with some backend work.  It follows in summary:

  • Refactored away the URL table.  Previously there had been a URL table which stored the urls as LONGTEXT objects and let the MANUSCRIPT and SAINT_SOURCE tables reference it by ID.  This was theoretically elegant, but kind of useless.  I replaced the URL_ID with a proper URL in both of those tables, mapped the data in and dropped the URL table.  This was done entirely in SQL, without a single line of application code needing editing, which perhaps made me more proud that it should
  • Initial work on a date-graph page is complete.  This gives a histogram of the dates of the DB contents, only for manuscripts with entries though.  It is not complete, but working and useful (see below)
  • Added the ability to add native css to the colors file.  This is outside the css_raw column in the COLORS table, that gets inserted into the named class.  Now there's a stand-alone CSS file that gets appended.  This was needed to implement different colors on the first letter of an entry, since those need pseudo-classes
  • Added a little secret tool to visualize ALL the manuscripts in the DB.  It's just a toy, but also only took about 45 minutes of coding

CoKLDB Date Distribution.png

Saints, now with Images

This addition has been a long time coming, and might be the last addition that adds a new type of content for a long time.  There is now a system for storing images and associating them with the saint data.  It consists of a series of AWS resources, principly an s3 bucket to store and serve images, and a lambda function to do automatic resizing, and a new table to store metadata.  A column was then added to the `saints` table to link to the `image` table.  All of this is accessed through a new class `Saint::Image`, which is created and stored as part of a `Saint` class.

To the browser, the change is only evident on the saint to manuscript popup or the saint detail page. The popup uses the small image, with description as a hover field, and the detail page uses the larger image. Examples of both are below

Saint popup image.png

saint detail image.png

From the collection of the Free Library of Philadelphia, currently being digitized as part of the BiblioPhilly project, comes Lewis E 90. This manuscript was written in Rouen for the Use of Coutances, about 200 km west on the Normandy coast, at the end of the 15th C. It is not a very attractive MS, the calendar borders are rough and generic, the miniatures functional, and the scribal hand irregular.

The calendar, however, is interesting for its accuracy and specificity. Out of 192 entries on 188 dates, all saints were identifiable from other sources and only 2 were off their proper date. Both of these, St. Wandregisil (1734) and St. Justus (3332) were displaced one day by a more important saint, Mary Magdalene (1730) and Luke the Evangelist (3333), respecively.

There are many rare observences recorded on the calendar, including several specific to Rouen. Five Rouennais bishops are named, along with two translations of relics to the city: St. Romain, for the Cathedral, on June 17 (2041) and an unammed saint on December 3 (4369).

(DB Id: 313)

Vat lat 9495, donated to the Biblioteca Apostolica Vaticana in the 19th C. by Pope Pius IX, is a rather late Book of Hours, made probably in the second quarter of the 16th century (1519-1550). Neither location of manufacture nor of use is listed, but due to several calendar entries, I would perhaps situate it in the area of Brussels.

The calendar has multiple unusual entries, some of which might point to Brussels. St. Gudile (1522) is indicated in red and her octave (4365) is not only indicated on Jan. 15 (see f.1r below), but has pushed St. Maurus (1253) to the following day. Less obviously useful for localization are several unique Octaves, St. Michael at Monte Gargano (4360), St. Catherine (4367), and St. Barbara (4368).

(DB Id: 243)

The first of a series of frontend/backend interface changes, jQuery loading has been implemented on the main manuscript list.  The reason for this is to give the user a better experience, the shell of the page loads instantly and the CGI code is called to populate the table.  While it is running, there's a solid loading screen provided by the jquery-loading plugin.  This feature will be extended to some of the other slow pages in the coming weeks. 

Take a look at the screenshot below or live.

list manuscripts loading.png

Two more from the Biblioteca Nacional de Espana.

Res/149, the Hours of Guillaume Rollin, was made in Valenciennes around 1460-70. The calendar is rather plain, with nothing but some penwork around the KL for decoration. It is, however, very accurate. Of 185 entries, only three are on the wrong date. Of the entries, there are no unique ones, but having Cornelius and Cyprianus (3696) on September 14th, in addition to the Exaltation of the Cross (3693), is an less common pairing.

(DB Id: 291)

Vitr/25/5, the Book of Hours Voustre Demeure, was made in Ghent circa 1472. As opposed to Res/149, above, the calendar here is nearly complete and heavily decorated. Each of the pages has illusionistic borders populated with insect, plants, and bugs. At the bottom of the versos, where the months start, there is a vignette of the labor of the month. On the rectos there is a scene with the zodiac in the sky over a seasonal landscape.

The entries in the calendar are less careful. It seems that the intent was to have a complete calendar, but in many months the insertion of the word “Nonas” (Mar., Apr., July, Oct., Nov., or Dec.) or “Idus” (Aug.) has filled the line and no saint exists for that day. There are only two actually empty dates, Oct. 26 and Dec. 14. As is common in complete calendars there are a fair number of unidentified saints and 30 on incorrect dates. One of the errors is St. Columba (4146), listed on Dec. 19. Though her feast should be on the 17th, there is one other manuscript in the DB, Walters W.195, with the same shifted saint. These are both Flemish MS from the late 15th century, and though there is no other known connection, this may indicate one.

(DB Id: 292)

Two more manuscripts from the Biblioteca Nacional de Espana, both Parisian, have been added in recent days.

Vitr/24/1 was in the collection of King Charles VIII (see 259), and dates to the very end of the 15th century. It has beautiful calendar, complete in French and graded gold, red/blue. Surrounding the labor of the month, on the outer margin, there are short French poems about the month. The entries are quite a mess, however. 43 appear to be on the wrong date with an additional 27 entries unidentified.
(DB Id: 287)

Vitr/24/7 is an earlier book, dating from circa 1400 with an unknown patron. It is named for a 19th century owner, the Duke of Osuna. The calendar is complete, with an unusual grading system, gold and red/black. Like the other Parisian Hours, this one has a calendar filled with errors. 52 entries are on the wrong date with an additional 65 unidentified.
(DB Id: 280)