Fortunately, it is rare for
metamodelers to go overboard with wonderful gradient ?¬?lls and ?¬‚ashy bitmaps at this
stage: clearly, there is no sense in polishing the symbols until the set of concepts is
stable. Basic symbols will suf?¬?ce for now.
There are, however, two things to avoid even at this early stage. Do not make all
concepts??™ symbols differ solely by color: the brain perceives such elements as
representing minor variations of the same thing. Similarly, there is hardly ever a need
to include the concept name as part of the symbol:DSMleverages the brain??™s ability to
recognize things as in the real world. Real-world tables and chairs do not come with
such labels, nor would the brain recognize them from the labels if they did (try
switching the labels if you don??™t believe us!).
13.4.3 Example Models and DSM Use Scenarios
A language is nothing until you use it to say something. Even invented languages like
Esperanto or Klingon stand or fall on whether there is anyone willing to speak them.
The textual programming languages we use today give good examples of languages
that have been designed well and had the rough corners knocked off them through
use??”and of languages where one or other of those processes has failed. As we saw in
Section 10.7, DSM languages are no exception: we need to test the language by
building example models with it.
Pages:
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645