[REBOL] [Bot] email command Bot
From: mauro::fontana::speedautomazione::it at: 26-Mar-2004 9:31
Hi all,
I have been developing a Bot that can receive emails on a specified email
address with a particular format, execute the commands and then answer
back the results.
I mainly thought it as where I work now I have not access to the web but
only to the corporate email address, but I wanted to be able to monitor
some things on my computer at home, and eventually, being able to make it
send me web pages through email.
I have lately seen that someone has uploaded a similar project on the
rebol.org library.
My code is far from being finished and, most of all, being in a neat form
to be made public.
Moreover, I am still working (quite slowly, I admit it) on the logging
feature that would allow me to create nice reports of all the info about
the requests sent to the Bot during time.
I would like to know if someone is interested in this project. I am quite
happy to send the code to anyone who would like to help me. The code is
OOP, that is the bot is a object with methods as is the logger. I would
like to add XML logging to it to be able to easily extract data later.
[small rant]
I am not a Rebol genius, though I have been using it (discontinued) for
some time now. There are still many things I have problems to understand,
mainly due to its "compactness" in the syntax, which I though it was a
thing the language wanted to avoid. See those neat examples in the library
to see what I mean. No doubts they are nice pieces of code that run well,
but they are quite ermetic (cryptic syntax)and often use some undocumented
parts of the language (like the use of internal ports, sub-ports, system
objects etc...) which make them quite "mysterious".
I think this is the first problem novices meet when starting grasping with
the language. The use of forms like :word 'word to-lit-word! to-lit-path!
set get etc... can be quite powerful but can also make the code a mess
(like doing things in C like c = *(unsinged short *) &value[]).
And often they are done just to keep the code short instead of making it
clear.
Then the fact that there's not exhaustive documentation on the system
objects and how to take advantage of them. I have still to understand how
the async:// thing works and where are those info about the
sub-ports/handlers explained in the docs. This is just an example.
Last a request, which may have already been asked, but I have not seen it.
Is it not possible to make the interpreter load binary code so that one
can "convert" the source code to it and have the interpreter load it not
in ASCII form? So, even though the reverse engineering of the binary could
be done, only the basic instructions are there and not the entire clear
source code. This way the users could create "embedded" applications
without having to mess with packages and source code distribution.
[/small rant]
Thanks for reading.
M&F
