On this page we will track changes to the Bridge and its data, as well as editorial decisions that have been made concerning the data.
- Summer 2018: Bridge algorithms completely rebuilt for speed and accuracy; added ability to create lists from non-contiguous sections of a work and/or multiple works; corrected problem with displaying proper names; corrected filter issues; added ability to see word count in section as well as text; added option to display “source-text” (text in which word appears when creating a list from multiple works); improved text selection mechanic on selection page (auto-completing input field); added contextual prompts for different texts (e.g. prompt for range of textbook has only one sub-division (1 to 2) but for poem with line data, it would prompt that those sub-divisions exist (1.1 to 1.5)); Include ability to filter entries with multiple declension/conjugation values (e.g. an adjective commonly used as a substantive and the display lemma includes both principal parts; N.B. data still needs to be improved to take full advantage of this); added ability to filter out CLTK stopwords on results page; innumerable minor improvements; significant curation of Latin and Greek dictionary entries and other morphological data; validation of all words in works (i.e. if we text data indicates that they appear in a work they will now display in lists); improved lemmatizer (advanced to beta); added: Claudian, Opera Omnia; Martial, Opera Omnia; Ps-Seneca, De Moribus or Proverbia, Romulus Anglicus Fabulae 1-10; Horace, Satires Book 2 (now have all), Selection from Maffeius’ Historiae Indicae; Seneca, De Ira; Seneca, De Constantia; CLTK stopwords
- Summer 2017: Add text specific definitions for selected texts; added a visual signal that the program is processing requests; adaptive display for mobile devices, corrected filter so that when filtering selection for parts of speech, activate category if a sub-category is checked (currently, the whole category must be active for sub-category to register); allow creation of similarity lists; Copy/save URL of output (to share, return to); ability to show first appearance and total appearances; sortable results table; significant increase in Latin texts; added many Greek texts; improved display of names; simplified selection mechanic (separate panes for text, lists, and textbooks).
- 12/29/2016: added Eutropius, Breviarium 1; corrected typo in lemma for VNDEVIGINTI and re uploaded Cambridge and JenneysRed to match new form.
- 12/28/2016: added Petronius, Satyricon 1-78; minor corrections to Latin dictionary.
- 12/12/2016: corrected SHORTDEF of ΙΣΤΗΜΙ; improvements to Greek results (still not complete)
- 12/7/2016: Horace, Odes 1.37 added; L’homond 20 (Gracchi) and 69 (Actium) added; Greek principal parts corrected (ἀπόλλυμι, ἐπστήσω; definition of εἰκός)
- 11/22/2016: Horace, Satires 1 uploaded; Cicero, In Catilinam I re-uploaded with corrected numeration
- 11/21/2016: Python script that concatenates locations of words in texts fixed.
Don’t yet see something that you’d like? Please leave a comment or contact Bret Mulligan.
Bridge 3.0 in-line improvements
- Allow user to select a threshold for results; e.g. show word only if it appears more than x times in text (avoid the “long-tail” of vocab; e.g. the 1782 words that only appear once in the Aeneid, the 845 that appear twice; the 484 that appear thrice, etc.; Ovid Amores 1: total words = 1540; 1x = 833; 2x = 288; 3x = 120; 4x = 72; 5x = 51; 6x = 47; 1-3x = 1241! or 80% of unique words occur less than 1-3 times).
- Set name of CSV download. At least append .tsv (or csv) to end; if it could be generated from the search string or a version of the URL, that would be even better.
- Include a title on printed lists.
- Add an exclude override feature that would highlight and/or NOT exclude a list of easily confused words (q- words, idioms, etc., EVEN if they would otherwise be excluded by the algorithm; neede: a list of such words (this would be a good things to have anyways)
- Add ability to exclude a Latin stop list.
- Fix first order of Appearence for texts with sub-sections (currently sorts 1.1, 1.10, 1.11., 1.12, 1.2, etc.)
- Allow for display of simultaneous display of multiple definition types (e.g. BASIC and TEXT-SPECFIC)
- Allow users to exclude a selection OR a frequency range for ancient texts. AND/OR include a frequency / appearances filter in filter pane
- Add text-specific principal parts (e.g. if you are reading Homer, you receive Homeric forms of principal parts; if Koine, Koine)
- Add multilingual support for application (native language front-end for select languages: e.g. French, Mandarin, Spanish)
- More granular data for verb morphology
- Add option to see cognates with definitions.
- Visual dictionary?
- Audio of Definitions
- Mobile/Web app to aid in vocabulary acquisition.