List Info

Thread: Re: DBmail webmail app




Re: DBmail webmail app
country flaguser name
United States
2007-06-07 06:47:25
Aaron Stone wrote:
> On Wed, Jun 6, 2007, Tom Allison <tomtacocat.net> said:
> 
>> On Jun 6, 2007, at 6:39 PM, Aaron Stone wrote:
>>
>>> On Wed, Jun 6, 2007, Tom Allison <tomtacocat.net> said:
>>>
>>>> Jake Anderson wrote:
>>>>> If you have 4 accounts then you won't
even notice it.
>>>> I don't think I will either and don't
believe that I ever said that.
>>>> All I'm trying to do is make sure we don't
attain the goal of  
>>>> handling 100MB
>>>> attachments regardless.  The ends never
justify the means.
>>> Meaning that you are proposing that we refuse
to allow 100MB  
>>> attachments
>>> in DBMail? With all due respect, that's silly.
>> Where did I ever say that?
>> I just can't find where I said we should refuse
attachments of any size.
> 
> Oh, cool, we're all on the same page then 
> Sorry for any misunderstanding of the thread!
> 

It took me a while to understand that:
A) some people use email for file storage vis-a-vis Drafts
(strange but true)
B) some day 100MB will seem typical just like Gates' 640KB
RAM comment...
C) within company sites the proliferation of a single file
attachment is large

It makes perfect sense to do this and it will probably only
make dbmail 
better/faster/more.

I'm also a big fan of KISS and strongly believe in the idea
that originated with 
UNIX programming (opposed to Windows) in that each program
should have a very 
distinct function and you can enhance features by making it
accessable rather 
than incorporating everything into it.

Examples of what I'm talking about are things like:

piping on the unix command line allows you to
"glue" features together very 
nicely and to rearrange the pieces at will.

Web Services/REST give you access to different functionality
through an open 
means of communication -- similar to piping but over a
network.

Agile et al likes lots of little simple functions that are
simple and specific. 
  Also known as DRY

The example of a blog and dbmail is a great point.  dbmail
isn't a blog server.

It's still a mail server and someone used the API of IMAP to
make it do 
something else.  This gives you the type of accessability
that makes your 
software rock -- but it's not dbmail's job to do blogging.

If you want to make dbmail capable of doing spam
filtering...  wouldn't it make 
more sense to simply pipe the spam filtering in front of the
dbmail interface 
like procmail does today?  I'm stuck on this one.  I don't
know where there is 
much advantage here.  I suppose you could give dbmail the
option to do 
redirection in the same fashion that postfix does
content_filter.  But then it's 
already been done and I still don't know what advantage
there is here.
I'm not going to say it's "bad" because it's not
for me to make that judgement. 
  But I can say "I don't get it" and can someone
provide me with a case where 
it's a "good thing" or better than what's
available today?

Personally, I would be more impressed if someone would fix
SIEVE so it could 
make external calls in the same fashion that
procmail/maildrop do.  Sieve is 
nice because it allows you to "stay off the disks"
by keeping all the rules and 
processing both in the database and per-user configurable. 
This makes virtual 
mail configurations much easier.
I don't know how to make global rules for Sieve so that's
"bad" in my book.  I 
also can't access an external process like I can with
procmail/maildrop.

Strictly my opinion, but if anything needed to be improved
it would be in sieve 
and not dbmail itself.  Unless dbmail wants to take on it's
own filtering rules 
and thereby develop a means of making these external
accesses.

I would love it if you could make rules that are in groups
of global, by domain, 
by user.  But I'm not sure I can do this.  Not really
tested.  Can you?  It was 
also be nice to make external calls like you can with
procmail (pipe in from 
dbmail, do something, pipe out to dbmail again).

This would allow you to very easily plug in your favorite
spam/virus filtering 
software.  But I would stress that it's a point of access
and not software 
inclusion.  It allows for the option/choice of the user.
_______________________________________________
DBmail mailing list
DBmaildbmail.org
htt
ps://mailman.fastxs.nl/mailman/listinfo/dbmail

Re: DBmail webmail app
country flaguser name
Austria
2007-06-07 17:41:12
On Donnerstag, 7. Juni 2007 Tom Allison wrote:
> If you want to make dbmail capable of doing spam
filtering...
>  wouldn't it make more sense to simply pipe the spam
filtering in
> front of the dbmail interface like procmail does today?
 I'm stuck on
> this one.  I don't know where there is much advantage
here.

It should never *do* spam filtering, but *support* it. The
idea is this:
- there are distributed checksums like DCC, pyzor, razor
- they are reported by thousands of mailservers, which
calculate 
checksums on every e-mail they receive
- once a certain checksum gets a lot of hits, the
probability that it's 
spam increases
- the disadvantage is, that if you're the first who gets hit
by a new 
spam wave, no checksums will exist
- so if you can, by the end of the day for example, later
reprocess 
those e-mails with a high probability of being spam, and
filter them, 
your system works better
- this is especially true for Sunday night, you can
reprocess those 
spams that arrived during the weekend, so people on monday
morning will 
have even less spam than before

It could well be that soon e-mails with pictures will be
delayed by an 
hour or so before giving them to the users, just to recheck
them for 
spam checksums later. It sounds strange now, but when spam
rates grow, 
things have to be done. Another option would be to reject
e-mails 
containing inline pictures that are highly probable spam,
and only 
allow attached pictures.

mfg zmi
-- 
// Michael Monnerie, Ing.BSc    -----      http://it-management.at
// Tel: 0676/846 914 666                     
.network.your.ideas.
// PGP Key:         "curl -s http://zmi.at/zmi.asc | gpg
--import"
// Fingerprint: EA39 8918 EDFF 0A68 ACFB  11B7 BA2D 060F
1C6F E6B0
// Keyserver: www.keyserver.net                   Key-ID:
1C6FE6B0

_______________________________________________
DBmail mailing list
DBmaildbmail.org
htt
ps://mailman.fastxs.nl/mailman/listinfo/dbmail

[1-2]

about | contact  Other archives ( Real Estate discussion Medical topics )