Comment to Interpreting MARC: Where’s the Bibliographic Data? by Jason Thomale Code4lib journal, Issue 11, 2010-09-21
100 1_ |a Bach, Johann Sebastian, |d 1685-1750.
240 10 |a Partitas, |m harpsichord, |n BWV 825, |r B major; |o arr.
245 10 |a Partita no. 1, BWV 825 |h [sound recording] ; |b Englische suite no. 3, BWV 808 = English suite = Suite anglaise ; Franzosische suite no. 2, BWV 813 = French suite = Suite francaise / |c Johann Sebastian Bach.
700 12 |a Bach, Johann Sebastian, |d 1685-1750. |t Englische Suiten. |n Nr. 3.
700 12 |a Bach, Johann Sebastian, |d 1685-1750. |t Franzosische Suiten, |n Nr. 2.
With our cataloging rules, the parsing you mention has always been performed manually, using separate 700 author/title analytic entries, plus the 240 field.
The fundamental purpose of the 245 field is not so much for search and retrieval, as to provide a reliable transcription of what appears on the title page, including all of the typos, etc. It is there for description and identification purposes.
The access of the item has always been through controlled fields. When keyword was introduced, while it added to access in certain ways, it also “threw a spanner in the works”, in other ways, when looking at it from a traditional viewpoint.
But yes, your basic point is correct: in many ways, MARC records are very akin to textual markup language.