I completed the article on "installing BP1", adding some background
information and links for readers who might have landed from another planet…
Reading again our paper on BP1 applied to the study of q‘aida
improvisation (https://halshs.archives-ouvertes.fr/halshs-00004506) I
remember the importance of early features of BP1, such as pattern
grammars (using a smart editor than could never be reproduced on the Mac
version), wildcards, negative context etc. These are available in BP2
but we have very few examples of their use. Grammars and data analyzed
in our 1992 paper are convincing examples but reading them requires
investing time to get familiar with the musical context.
Nonetheless, I am confident that we can publish an article on the
website taking grammar "Q2.7" (from disk BP-JULY87.DSK) which is the
first one described in our paper. The aim would be to demonstrate:
1) Producing variations with the grammar;
2) Analyzing data examples with the grammar (where templates are used);
3) Assigning weights to grammar rules from a set of (correct) examples —
our first attempt of "machine learning".
The same can be done using BP2 since all algorithms have been
reproduced. Trying them on BP1 might give new ideas about their proper
implementation in the next version.
Bernard Bel
---------
Hello BP1 fan club!
I started documenting the installation and basic operation of BP1 in its
Apple II emulation:
https://bolprocessor.org/installing-bp1/
It is actually very easy to install but I wanted to describe the process
in detail. A trial with LinApple on Linux would be welcome… I had done
the test on Windows in Marseille.
Don't hesitate to comment, suggest clarifications and fix language
mistakes! (Anthony can edit the post directly.)
I start remembering how it worked ;-)
The next publication will be an explanation of its use, preferably using
musical examples which had been well documented and mentioned in
publications. For this reason I selected material dating to 1988, our
latest work on this machine…
Bernard
|