The release contains many changes for ease of use and functional extensions:
Support for the Java 1.0 API was discontinued, (thus e.g. JDK1.0.2, Internet Explorer 3.0 and Netscape Navigator 3.0 are not sufficient for MINERVA.)
MINERVA now accepts the parameter -syntax strict|classic to use unicode character mapping vs ascii-only for reading.
open/3/4 accepts an option encoding(Encoding) to open individual files in the specified encoding. Encoding is e.g. 8859_1, EUCJIS, SJIS, UnicodeLittle etc for ascii, euc, sjis, and little-endian unicode respectively.
open/3/4 accepts an option syntax(Syntax) to open individual files in the specified encoding. Syntax is strict or classic for unicode mapping or ascii-only mapping of characters.
For *.mpk MINERVA package files their encoding can be specified in the using *.msp MINERVA Startup Properties file with entries encoding.Interface=Encoding like:
file.system.mpk=stdlib/system.mpk file.gui.mpk=gui/gui.mpk encoding.system.mpk=8859_1 encoding.guibuilder.mpk=8859_1
foo(x,2). ?- N is foo(x) + 1.
load('ftp://machine/file.mbc')
load('http://machine/file.mbc')
load('directory/file.mbc')
Examples:
minervar minerva minervar myservice
Such a bundle may contain various files and is primarily intended for *.mbc MINERVA binary code files. It should contain one package package.main. This will be loaded if no file (-l option) with a main/1 definition is specified. It is expected that this package contains a predicate main/1 which will serve as the entry to the application.
MINERVA knows the parameter -c Name, used to announce a *.mca file to MINERVA. You should be able to notice the advantage mainly when loading over the net.
As a result, you can now write MINERVA applets that require only two HTTP GET operations, thus dramatically decreasing loading times over networked connections.
Example:
get_property('license.licensee', Owner).
Objects can be created with constructors, arrays with create_array/3.
| scroll to top |
|