[clean-list] Re: File Reader Combinators

Martijn Vervoort martijnv@cs.kun.nl
Mon, 27 Aug 2001 12:17:25 +0200


Dear readers,

> Two reasons for no pursuing the File Reader Combinators too far, were:
> 1) Dynamics coming up. They will enable me to write and read anything
without any programming effort.
> 2) They open up a huge area of design decisions to be made. The fact that
they can always be made more beautiful and general, has hindered me going
forward fast to make something useful and complete.
>
> Three reasons to resume work on them:
> 1) The implementation of Dynamics seem to suffer quite a delay.

True. Compiler support is relatively simple compared to the support needed
from the {static,dynamic}-run-time system.

> 2) Dynamics will probably only able to read files that were created by
writing a Dynamic. I don't expect Dynamics to use some XML, EDIFACT, or any
other standard format.

> 3) What I read about Dynamics did not mention error checking and
reporting. What if a 10 MB Dynamic has suffered some crippling somewhere?
Will I get some details about the error? If not from the regular dynamic
reader function, will some utility-program be able to
analyze Dynamics?

There exist conversion functions from a dynamic to a string and vice versa.
So if you need to put and extract the encoded dynamic
in and from the format you want to use.

Error handling is supported but only minimally at the language level.

At the moment there is a utility to show dynamics which also does some error
checking. This utility will probably be extended to do
extensive error checking.

Kind regards,
Martijn Vervoort