Jargons of Info Tech industry

[Chris Head]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

John Bokma wrote:
> Chris Head <chris2k01 at hotmail.com> wrote:
> 
> 
>>John Bokma wrote:
>>
>>>Chris Head <chris2k01 at hotmail.com> wrote:
> 
> 
> [HTML]
> 
> 
>>>It can be made much faster. There will always be a delay since
>>>messages have to be downloaded, but with a fast connection and a good
>>>design, the delay will be very very small and the advantages are big.
>>
>>What advantages would those be (other than access from 'net cafes, but
>>see below)?
> 
> 
> And workplaces. Some people have more then one computer in the house. My 
> partner can check her email when I had her over the computer. When I 
> want to check my email when she is using it, I have to change the 
> session, fire up Thunderbird (which eats away 20M), and change the 
> session back.
> 
> [ .. ]

Hmm. That would just be a matter of preference. Personally I moved my
Thunderbird profile into a shared directory and pointed everyone at it.
Now only one login session can run Thunderbird at a time, but any login
can see everyone's mailboxes.

> 
> 
>>>Each has it's place. A bug in a thick client means each and everyone
>>>has to be fixed. With a thin one, just one has to be fixed :-D. 
>>
>>True. However, if people are annoyed by a Thunderbird bug, once it's
>>fixed, most people will probably go and download the fix (the
>>Thunderbird developers really only need to fix the bug once too).
> 
> 
> Most people who use Thunderbird, yes. Different with OE, I am sure. With 
> a thin client *everybody*.

True. As a programmer I don't usually think about the people who never
download updates. The way I look at it, if somebody doesn't have the
latest version, they shouldn't be complaining about a bug. I guess thin
clients could be taken to mean you have a very light-weight auto-update
system ;)

> 
> 
>>>Depends on where your mailbox resides. Isn't there something called
>>>MAPI? (I haven't used it myself, but I recall something like that). 
>>
>>IMAP. It stores the messages on the server. Even so, it only has to
>>transfer the messages, not the bloated UI.
> 
> 
> But technically the UI (whether bloated or not) can be cached, and with 
> Ajax/Frames, etc. there is not really a need to refresh the entire page. 
> With smarter techniques (like automatically zipping pages), and 
> techniques like transmitting only deltas (Google experimented with this 
> some time ago) and better and faster rendering, the UI could be as fast 
> as a normal UI. 
> 
> Isn't the UI in Thunderbird and Firefox created using JavaScript and 
> XML? Isn't how future UIs are going to be made?

I believe it is. I'm not sure if it's a good idea, but that's neither
here nor there.

> 
> 
>>I concede that Webmail
>>might be just as fast when using a perfectly-designed
>>Javascript/frames-driven interface. In the real world, Webmail isn't
>>(unfortunately) that perfect.
> 
> 
> Maybe because a lot of users aren't really heavy users. A nice example 
> (IMO) of a web client that works quite good: webmessenger ( 
> http://webmessenger.msn.com/ ). It has been some time since I used it 
> the last time, but if I recall correctly I hardly noticed that I was 
> chatting in a JavaScript pop up window.

Haven't ever needed to use that program.

> 
> 
>>As I said above regarding 'net cafes:
>>
>>If the Internet cafe has an e-mail client installed on their
>>computers, you could use IMAP to access your messages. You'd have to
>>do a bit more configuration than for Webmail, so it depends on the
>>user I guess. Personally I doubt my ISP would like me saving a few
>>hundred megs of e-mail on their server, while Thunderbird is quite
>>happy to have 1504 messages in my Inbox on my local machine. If I had
>>to use an Internet cafe, I would rather use IMAP than Webmail.
> 
> 
> I rather have my email stored locally :-) But several webmail services 
> offer a form to download email.

I've not seen a service that allows that. Sounds nice.

> 
> 
>>>>Ergo,
>>>>Thunderbird is faster as soon as the Internet gets congested.
>>>
>>>Ah, yeah, wasn't that predicted to happen in like 2001?
>>
>>Wasn't what predicted to happen? Congestion? It happens even today
>>(maybe it's the Internet, maybe it's the server, whatever...). Hotmail
>>is often pretty slow.
> 
> 
> I read sometime ago that about 1/3 of traffic consists out of bittorrent 
> traffic... If the Internet gets congested, new techniques are needed, 
> like mod_gzip on every server, a way to transfer only deltas of webpages 
> if an update occured (like Google did some time ago). Better handling of 
> RSS (I have the impression that there is no "page has not been 
> modified" thing like with HTML, or at least I see quite some clients 
> fetch my feed every hour, again and again).
> 

Eventually you reach the point where it's not bandwidth any more, it's
server load. All these things like mod_gzip, deltas, and so on add
server load.

As to the point about "page not modified", it's not in the HTML spec,
it's in the HTTP spec. RFC2616 (HTTP1.1) defines an "If-Modified-Since"
header a client may send to the server indicating that it has a cached
copy of the page at that date. If the page has not changed, the server
should send HTTP 304 (not modified) with no content. For best results
(due to clock mismatches etc), the client should set the
If-Modified-Since header to the value of the Last-Modified header sent
by the server when the page was first requested and cached.

I think we can agree that in some cases, Webmail is better, and in
others, clients are better. Much of this will be personal preference,
and I would like to see ISPs offering both methods of accessing e-mail
(as mine in fact does - POP3 and Webmail).

Chris
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (MingW32)

iD8DBQFDELff6ZGQ8LKA8nwRApxiAKDBU2R5KYAhp/4MJDoLlrbC5hWpLgCeNbnh
YK2tCasrMOY3SaUV1gMtZdg=
=N2wk
-----END PGP SIGNATURE-----