[Webfunds-devel] wizards & warlocks

Wed, 16 Aug 2000 21:21:42 -0400

> >> Save button is a bad idea, as it defeats the idea of a wizard, but yes it
> >> is a good idea to have a similar feature.
> >
> > This was indeed gist of conversation; that a Wizard does things as it does
> > things and changing it is not a Good Thing (tm).
> >
> > My comment on that is that I need these features to make the code
> > usable and workable in the face of users.  If the user interface
> > for a Wizard is unchangeable & A Bad Thing (tm) then I'm quite happy
> > if the changes are made and we change the name to Warlock, Witch,
> > or any other Wonderous Thing (tm).
> >
> > The problem with not having the Save button (and other non-Wizard
> > features) is that I and other contract writers have to go through this
> > dozens of times before we get it right, so we end up walking thru
> > the process so many times that we tear our hair out over doing the
> > same 3 minute typing process over and over and over....
> 
> No save button does not mean that it cannot be saved, it just means that it
> has to be implemented a bit different in a wizard: not with a save button,
> but with a box asking for the filename and then saving on clicking 'next'.
> In this case this would just be an extra box on the FinishSig panel.
> 
> The main advantage of a wizard is that it can be used by anyone without
> studying any documentation. The disadvantage is that advanced users find
> wizards annoying. But if you only want to target advanced users, why not
> write a command line tool that does the same thing in minimal time, once
> you understand what the f*ck it is doing?

Well, somewhere in between.  I need a "wizard" that can be used by
'advanced' users.  Which is why I'm quite happy to call it something
else.  Or, to put it another way, the requirement for a wizard was
fundamentally incorrect, we need something that is more advanced
than a wizard, whatever that may be.

Preparing contracts is a difficult task.  There are many things to
check.  Doing it via command line is a possibility, but impractical
because contract-preparing people don't normally know what a command
line is.  In fact, they aren't in general so clued up as to what a
computer is ... we're talking lawyers, doctors here an exception,
Unix bufs are definately out of place.

The issue here is that contract preparation is not a one-off serial
task.  It *is* mostly a linear process, but it is a line that is
travelled many times, backwards and forwards, and restarted many
times.  So maybe this thing is more of a process editor.

My image of the perfect process - always assuming infinite budget and
resource - is one where the proto-contract is built up from nothing,
saved and restored, filling in the dialogs with its contents, and
gradually as time and new info is built up, proceeding to a final
signing.  There are even many preliminary signings, it is indeed only
when a contract is issued to the server, minted by the mintor and
finally distributed as value to innocent users that we cross a point
of no return.

The contract warlock (!) is part of that process. Anything it can do
to assist this is warmly appreciated.

As an aside, the cost of issuance of one of these style of instruments
is about $600k on the NASDAQ - the cheap US market - and GBP 600k on
the LSE.  We are looking to knock zeroes - more than one - off those
numbers, so it should be no surprise if it is both more complex than
we thought, and more powerful than we dreamed once we get it done :-)

-- 
iang