po blog — po pro


Listen to someone speak. Listen to their said. Listen to sound. Listen to inflection, expression. Listen to colour, rhythm, pitch. Hear context, semantics, the meaning. Hear information, expression: hear the words.

So you’d expect word processors to process this; the sounds, the semantics, the inflection. None do. No word processor processes words. Tell me otherwise, tell me which regards how sound aligns. Tell me which rebound in rhythm. Those so–called word processors don’t process words, they process documents.

Document processor are fine, as such. What’s wrong is the misrepresentation, the implication of words; all of words, everything about words. They do not process words. They processes letters, symbols in rectangles, no more. These so–called word processors process words like old buggers bugger the old.

PoPro will be software that processes words. It will use written symbols, it will process those symbols’ sounds, patterns of sounds; it will process semantics, inflection and across; PoPro will be the world’s first word processor.

And here’s where it’s planted. Right now it’s in design.


The design is documented separately, in requirements and architecture.


The code will be written in portable C++. It will be multi–threaded, to support, as much as possible, multi–core CPUs. Coding standards will approach Alexandrescu & Sutter, & use C++ fully. The code will use boost and TR1; although TR1 is currently very new, by the time this product is ready, the major library constructors (I’m aware of boost, Dinkumware and GNU) should have reduced any serious implementation problems.

Questions to consider:

  • Does portability require the GNU tools?
  • Which compilers? At least Visual C++ and GNU?
  • Which develop environments? Visual Studio, definitely, but what about Unix? Will the Mac environment properly support developing portable applications? Is there a sophisticated Linux environment? I didn’t find one when I last looked.


At this stage, testing consists of looking for flaws in the design. This is not rigorous.

Testing must be designed into the product so the product’s functionality can be tested automatically. A test environment must be prepared to conduct those tests automatically.


If no one else comes up with a fairly obvious idea like po pro, the chances are no one else believes it can work. I very much doubt it would work as only buy–in–the–shop commercial software, for example. Fortunately, commercial success is not my primary goal, although the pocket money would be nice. Anyway, that means there’s no real risk in publishing it.

If there is something original in it, by publishing it online it might be regarded as prior art. As I understand it, this would avoid the proposed legal system’s tax on innovation, presuming software patents are enacted in the EU. I was quoted a legal tax of at least UKP150,000 for po pro; had software patents been in place I’d be unable to proceed because I have absolutely nothing like that kind of money, and certainly would have no intention of trying to raise it. I know software patents are meant to protect innovation; it’s just that, in a field like software where no investment is necessary to innovate, so there’s no investment to protect, all they actually protect is corporate and patent lawyers’ opportunism.

ancient front