[fpc-other] Last message about changes (guaranteed!)

[Cephas Atheos]

I'd like to offer some fairly straight-shooting thoughts about the
communication and site interaction issues, separately to the idea of a
static site with a mail back-end.

So let me look at the way things are now, and the way they could become, and
give some realistic ideas of what it would cost, and what it would take, to
make such a significant change, based solely on what you've described so
far. Note that I'm not even thinking about software distribution just yet,
only the way the community operates and communicates, as I'm not qualified
to even think about software distribution.

One thing I do need to reiterate is that I'm addressing this whole issue
from the aspect of a brand-new, modern-day user, familiar with current
internet concepts.

The vast majority of these users don't know very much (if anything) about
svn, mailing lists mechanisms, newsgroups, or any of the technologies that
we take for granted. Have a look at the way Yahoo and Google implement their
user interactions =AD they universally implement not newsgroups or nntp
clients, not mailinglist access, but a web interface, without exception.
Unfortunately, that's the way of the world these days, and will be so for
some time to come =AD until the Next Big Thing happens along, I guess!
Basically, the sites that rely on nntp access are niche sites with fixed
audiences and known demographics, and even they could do much better if they
tried.

I do understand and appreciate that you're volunteers, and your budget
doesn't extend to multiply-redundant mainframes or dedicated systems
managers. That's true of me, too! And I would never intentionally suggest an
expensive or high-maintenance solution, ever. Nor would I suggest tools that
are the same, or worse, than what you're using right now.

I'm really surprised, to be perfectly honest, at the bizarre perceptions of
forum software expressed in the last few days here in these groups!

Forums don't actually work the way some of you think they do, I swear! I
would be interested to hear some examples of the terrible forums you've
mentioned at various times, they must be shockers. Of course, I don't
consider the string-and-glue forums, like Joomla or the pre-canned template
type sites as forums, they're more interactive html, which doesn't meet
anyone's needs.

I regularly (at least weekly) interact with (at current count) 74 forums and
another 25 or so other sites (wikis, ftp, and so on). Most of these forums
are PHPBB-based, to some degree, and the rest are similar, though coded
differently. So I know what I'm talking about with forum interactions and
what works and what doesn't. The best types of forums are those that allow
users to be users of the respective software or hobby, without needing to
figure out how to post a message, how to reply, and so on. NNTP sites are at
the bottom of that list, needless to say.

Without exaggerating or belittling nntp and smtp clients and servers, a
forum would do everything =AD and I quite literally mean EVERYTHING =AD tha=
t the
very best nntp/smtp software does, and it will do it out-of-the-box.
Literally.

I appreciate that you don't want to have to rely on a single point of
support or failure, and you really don't want to have to pay someone else
for doing what you can probably do already now!

But is it more difficult to manage a single forum, than it is to manage what
you're using now?

How many of you reading this message could, say, at 6:30 tonight, fix a
configuration problem on the mailing list server, if whoever usually looks
after it got called away and was uncontactable? From where you sit right
now? What about if a hacker logs in to the list and starts spamming the
list? How long would it take to shut them down? Would it be possible to keep
the list running? What if they got in all the FPC lists? How long would that
take? And how long would the server be out of action while that was
happening?

I've been there and done that, so I can imagine the chaos - unless you all
know each other's phone numbers and you all know how to configure the list
server so you can find out who the person is, shut them down, and lock them
out, then find a way to identify and remove the messages, then bring up the
server again, and so on.

On a properly set up forum, any one of you could log in, disable the
account, remove all the messages, and prevent the user from logging in
again, in less than a couple of minutes. From your iPhone.

While you're on the bus.

Tell me you can do that with your current software, and I'll shut up about
forums for ever! :) And I haven't mentioned the website =AD I can only guess
at how difficult it is to arrange for a new release of software or even a
news update.

So there are some major, and many minor, benefits to looking seriously at a
good forum installation to take the FPC project forward, just from the
perspective of keeping things running, and minimising hassle when something
does go wrong!

So what about the cost of the software?

The kind of interface I'm thinking of consists of open-source software,
exclusively. That's apache, PHP, MySQL, and sendmail. These are not only far
more robust and secure than news and mail list servers, they're incredibly
easy to install, configure, and maintain. If a 10-year old kid can run a
busy (1,500+ messages-per-day) forum after school hours, there's no reason
why we can't get something working that will delight and impress the pants
off all of us in our little corner of the web. Consider also that this would
allow remote administration via web interface from anywhere and any device.
All you need are the login credentials, and I doubt you'll be handing those
around willy-nilly! So security (in terms of management access) is no more
difficult than what you're using now.

So realistically, it would be a huge improvement to make the fpc project
available from the same server, at little or no cost for the upgrade, with
solid, secure, safe open-source components, presenting an easy-to-use,
easy-to-manage access portal that everyone here will know how to use without
reading readme's or compiling a single line of code.

Every new user would know instantly how to log in, how to search, and how to
ask questions and how to find answers, with no change that anyone has to
make to what they already have on all of our computers. If you've got a web
browser, you've got secure access. Who wants to search for source code for
an nntp client, that may or may not actually work with the server's nntp
server, and still have to verify their internal email client's settings to
make sure they can both send and receive messages? Who wants to wait until
the end of the day for a long list of messages about subjects you're
probably mostly not the slightest bit interested in, looking for a response
to a question you've posted the day before? Not me, and not anyone I know
of. I'd rather get immediate notifications of threads (questions) I'm
involved with, or interested in, than wait 24 hours for any possible
response!

Hey, if you like pain, go for it! But please don't expect new users to want
to go through the same arcane process. Because I can tell you, right now,
that they don't. And they aren't going to in the future, either.

This still leaves a lot to discuss. We need to understand some pretty
critical information still. Like how big are your mailing archives? Judging
by the statistics on the FPC main page, you have a grand total of around 700
questions, in total, since 2000!!

Of course I'm assuming that those statistics are wrong (by at least a couple
of orders of magnitude!), so it would help if we could find out the actual
size, in terms of storage capacity and number of messages. If you've got a
typical mail list archive, we'd need to think about how to migrate 50,000 or
so individual messages from each year into the database, which would imply
that we need pretty robust automated tools.

But if there are fewer messages than is usual for a project of this size and
complexity, it might make more sense just to write some shell scripts or
something. Or maybe an FPC project competition? Get the users interested and
involved and show off their skills writing platform-independent code to
retrieve and reindex and store the messages on a remote database server?

I wanted to say more, but I've already made this painfully long. I'm sorry
about the length, but these are ideas and concepts that would better be
discussed face-to-face, where we would have the luxury of telling me to shut
the **** up!! ;) So I'd rather take a long time to explain exactly what I
mean, than risk 2 or 3 or more days of back-and-forth "What do you mean by
----?""But have you thought about -----?" and so on.

Please have a think about all this, and please don't kill me for posting
such a humungous glob of a message yet again!

Hopefully, these are ideas many of you have already been considering for
some time, and if not, I do hope they give you some hope for new ideas in
the future!

Again, I'll leave the subject in your hands now, I've been harassing you
enough, and you've already been so patient and helpful and interesting!

-Pete

P.S. Sven, if you're referring to my spelling using the letter 's', that's
because I'm Australian, and we use the British system of spelling words
ending in =ADize with an 's'. Hopefully that explains what you wanted to kn=
ow!
Otherwise, I'm afraid I have no idea what you mean by the "S" question you
posed, but please clarify if you need! Hope this helps! -Pete










-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.freepascal.org/lists/fpc-other/attachments/20120929/a030c=
638/attachment.htm