At 11:46 AM 1/19/00 -0600, [EMAIL PROTECTED] wrote:
>
> Often dynamic scoping gets used as a way to provide
> "hidden arguments" to a function without the coding
> effort (or possible r/t overhead) of explicitly
> passing them as parameters.
>
> One way to get around this (and to reduce the consequent
> side-effect dependence and subtle bug opportunities) is
> to restructure the code so that all arguments are
> explicitly passed, instead of hidden in modifications to
> a global environment.
>
>Would this sort of additional commentary be of value?
yes it would and yes that was the reason -
the shadow was used to disable some
feature without passing the variable
explicitly from layer-to-layer.
more specifically, to port the code
to the new language quicker
(faster adoption) I occasionally
have to mimic a bug even if redesign
or re-implement with more explicit
state/context would have been better.
often, the hired-gun is not compensated for
redesign or redesign would steal thunder from
the original architect. politically a bad move
if the architect is still president.
changing language does not offend folks
at certain levels, making known bugs
go away brings up fears that behavior
is changed and a whole separate product line
is being created.
gotta be carefull to limit the number of
battles taken on at once.
An existing corporate infrastructure can only absorb
4+-2 changes per 'season'. note that the
number is smaller than the usual 5+-2 reported
for singular humans. in the corporate metaphor,
change is not good. most money is made via 'exclusive'
agreements, licensing, "terms", ROI and other things
that mean a stable/bankable position.
first problem to solve is 'sponsorship'.
no sponsorship means steep uphill battle.
(uphill: many years, plenty bad blood, little fun)
second is to retain sponsorship.
sponsorship is quickly drawn away
if support requirements have potential for going up.
convice them (management) the move changes nothing
in behaviors therefore no changes in support
requirements. the better the guarantee here
the higher probability you will be allowed
into the 'third-stage' of your 'master-plan'.
[see thread on "how do I institute cange to rebol"
for continuation and a sucessful master-plan
executed by a now well known author/language]
{-----}
[EMAIL PROTECTED]
