Registered: Jan, 2007
Re: Software Development Has Stalled
Posted: Feb 12, 2010 12:12 PM
> A couple of points flow (forgive the pun) from this:
> - in commercial/business applications, unless and until
> there is something, semantically, to replace that
> requisition form, all we're doing is scribbling around the
> edges. Even Apple, for all its alleged innovation, is
> still implementing the PARC interface. How old is that?
> About 1980, depending on how you measure. Consider the
> e names of our standard widgets: list, button, radio
> button, check box, tab, menu, and so on. All just
> pixelated analogs of paper, pencil, and mechanical toys.
> It's still just a paper form asking for some stuff.
> . We've not really innovated, just adapted. The the
> light pen/screen, and mouse, that was innovation. But
> that requisition form had already been adapted to the
> character mode VT-100 (and some would argue in a superior
Agreed. Essentially, the invariant in this scenario is the data - set of data elements to be specific - while the variants are the persistence and transport mechanisms. To continue my dream of being a broken record, company leaders already have digital persistence and transport. What can we now entice them with?
> - replace forms, with what?
This question is critical to ask.
> It just could be that the
> requisition form really is the best vehicle for asking for
> some stuff.
If by requisition form you mean the data elements collected that are necessary for decision-making on a potential purchase order, then absolutely yes.
> Limiting ourselves to existing hardware, how
> can we make it easier?
Bingo! The question is, what can we make easier? Electronic transport and persistence of the "form" are already implemented.
> My contention has always been that
> GUI's are maximally useful (or even minimally useful, when
> I'm in a bad mood) only when you never touch the keyboard:
> all input is picked with the mouse/whatever.
I'm of the same mind.
> In order to
> o do that, we need to present all input in limited
> options. In order to do that, we need to disaggregate
> data structures. You see where this is headed? BCNF
> databases; narrow "records", easily displayed (list,
> buttons, boxes) and touched. But that flies in the face
> of flat-file zealots from xml-land. With increasing
> density of screens, it becomes easier to do, if the coders
> would only be compelled to let go of datastore
I'm extremely comfortable thinking in sets. With XML, I'm still experimenting to see where its potential is.