At 2:01 PM -0800 on 12/12/99, Alain Farmer wrote:
>> Alain: I knew you were away for a while, Adrian.
>> Had I been a little less busy, it might have
>occurred
>> to me to notify you by personal E-mail that a
>FreeCard
>> vote was in progress. Perhaps we should include in
>our
>> voting process a personal-notification of all
>members
>> among the lessons learned with this first vote.
>
>Anthony: This would definitely be a nice feature to
>add to the voting CGI -- whenever a vote is started, a
>form should be sent to the list members which they
>just need to fill out and send back.
I don't think I said that. Don't want to steal credit for someone else's
great idea.
>Anthony: This is rather hard without running your own
>mailserver.
>
>Alain: Not necessarily. Having HyperCard 2.4 handle a
>mailTo URL is a cinch.
Yes, HC can send mail. But it's a lot harder to make it recieve mail. The
idea here would be the CGI would send out a ballot, and then people return
the ballot by email or www -- however they want.
The CGI would autocount ballots.
>So, what you propose is definitely
>possible, but I would like you to tell me why I should
>go to all of that "trouble".
Well, first off, the Linux server will kick as Mac's a** when it comes to
speed and reliability. We won't have to worry about ANY cgi bringing down
the box, no matter how buggy. Also, due to real file permissions, we won't
have to worry about someone compromising the server should we give them a
login. And running ftp, telnet, etc. will be much easier. Telnet will be
great for remote builds. I should also mention RCS & CVS! Real version
control. People able to check out the latest sources, modify them, check
them in, etc. Perfect for when we have a larger developer community.
Second, while the sending and autocounting of votes could be done without a
unix-varient (such as Linux) as long as you have your own mailserver, it
would be a lot harder. With Linux, all the program has to do is be called
when mail arives (sort of like a real Majordomo) and then process the
message. On a Mac, the program would have to continuously poll for new mail
by opening a POP3 connection to the mailserver, checking for new mail,
downloading any there, then processing.
The Linux solution would be much faster, and much more prone to work. It
couldn't be written in xTalk (at least not yet), but rather perl, I'd guess.
