diff options
Diffstat (limited to 'README.txt')
| -rw-r--r-- | README.txt | 45 |
1 files changed, 45 insertions, 0 deletions
diff --git a/README.txt b/README.txt new file mode 100644 index 0000000..b715526 --- /dev/null +++ b/README.txt @@ -0,0 +1,45 @@ + README for pd external package 'pdstring' + + Last updated for pdstring v0.02 + +DESCRIPTION + The 'pdstring' package contains objects for converting to and from + (ASCII)-strings, represented as lists of floats. + +INSTALLATION + Issue the following commands to the shell: + + cd PACKAGE-XX.YY (or wherever you extracted the distribution) + ./configure + make + make install + +BUILD OPTIONS + The 'configure' script supports the following options, among others: + + * --enable-debug , --disable-debug + Whether to enable verbose debugging messages. Default=no. + +ACKNOWLEDGEMENTS + PD by Miller Puckette and others. + + Ideas, black magic, and other nuggets of information drawn from code by + Guenter Geiger, Larry Troxler, and iohannes m zmoelnig. + +KNOWN BUGS + Memory Usage + Encoding each byte of a string as its own float is shamefully wasteful: + it uses only 1 byte out of at least 3 which could be losslessly used + given ANSI/IEEE Std 754-1985 floats, not to mention the remaining + byte(s) (usually 1) of the float itself or the (usually 4) bytes used + for the a_type flag. Unfortunately, Pd trims some floating point + precision in message boxes and in float atoms, so a truly lossless float + encoding for Pd would only be possible using 2 bytes per float (wasting + 1/2 the space of the float itself), and (to me), the memory saving such + an encoding would provide is just not worth the lack of transparency and + additional workload it would involve (but contact me if you want the + code anyways). + +AUTHOR + Bryan Jurish <moocow@ling.uni-potsdam.de> + |