Top Posting and Mobiles (OT From Re: [jdev] XMPP Ping/Keepalive: Recommended method ?)

Dave Cridland dave at
Fri Jun 23 04:05:25 CDT 2006

On Fri Jun 23 01:47:38 2006, ennova2005-jabber at wrote:
> p.s  Answering below quoted text  is similarly annoying for people 
> with mobile email clients where only the first X bytes are 
> retrieved by default ;-)

Be careful about making arguments like this. There's always the risk 
that a bored mobile email researcher will be reading, and want to 
deconstruct your argument and analyze it. :-)

The implication of this argument is that the primary saving important 
to mobile email is data transfer - this is certainly a key issue, but 
not the only one. Round-trips are also considered critical - more so 
in many cases, because the round-trip times on, say, GPRS, can rise 
from 700ms to 2100ms, which is quite obviously well within human 

But it's fair to say that downloading excess body part data is to be 

Now, with top-posting, in general the entirety of the message is 
appended to the end of the text itself. This is distinctly 
problematic, because the mobile handset (or laptop on GPRS, or 770, 
or whatever) cannot know when the message ends, and the original 
message starts. This actually leads to the behaviour you describe, 
because much of the message data *is* redundant. On the other hand, 
it's a guessing game - the client might not download enough data - 
causing an extra round-trip, or might download too much - causing 
excess download. In general, reading the entirety of the new message 
will end up downloading some of the original message.

Ideally, you could just drop the whole of the original message. Email 
has sufficient metadata that should the user wish to refer back to 
the original message, navigation to the In-Reply-To message-id is 
reasonably easy. However, this has additional, human related problems 
- it means that the author of the message has to write considerably 
more verbosely, and with a much higher clarity, in order to answer a 
message, and the reader has to recall the original message quite 
clearly to avoid having to download that.

Instead, then, the typical author can simply cite, or quote, small 
excerpts from the originating message, and reply directly to them, 
one point at a time. The familiarity of the orignal text will allow 
the reader to handle the new content faster, and the author to be 
more concise.

As a side-benefit, when reading through mailing lists on a compressed 
channel, the quoted portions compress very nicely relative to the 
previous messages, thus lowering the bandwidth cost even more.

Incidentally, Maciek's message also has problems for mobile clients, 
because it doesn't use format=flowed, meaning users with especially 
small screen sizes cannot easily distinguish between the quoted 
portions of the text, and his own additions. Also, the inline PGP 
signature (rather than using MIME) causes additional wasted download, 
although not very much. (Your top posting accounts for much more).

So to summarize:
1) Top posting is bad on mobiles, too.
2) Format=flowed is good for humans and computers.
3) Using MIME to keep content seperate is good. (This includes if 
there's a good reason to include the original message, in which case 
include it as an attachment.)

Dave Cridland - mailto:dave at - xmpp:dwd at
  - acap://
Infotrope Polymer - ACAP, IMAP, ESMTP, and Lemonade

More information about the JDev mailing list