yes taglal , otherwise i wouldn't have bothered my self reverse-engineering whatsapp and bought a cheap andro' to use whatsapp . i find it no use to have whatsapp as an APP on N900 , when we have something like telepathy which is easy to access and manage via a click and it's always running and wouldn't need much ram/cpu as a stand alone APP would .
your app might be closed by mistake , forget about you have to find your contacts from within the app instead of having them laying around on the desktop or the contacts list .
running a stand alone whatsapp app will require more used ram and cpu(means less battery time) and app that might malfunction on connection change 3g disconnects while travelling , non default IM tones/notfications/options ...
oops you restared and forgot to start the app , you want to go offline form all acounts ... wheres that app ...
dude i prefer purple plugin or telepathy or whatever it is ! at least you don't have to write that much code for a stand alone app (unless if you're aim is to collect donations or sell it ?) . at least telepathy already has xmpp protocol , so you just have to write the login/verification/registration part in a plugin for the N900 communication center -nice name haha-

dude, you are totally missing the point.

oh, i think teamer's point of view and attitude is perfectly correct. There's a "doing it the right way" and a "let's put in only the effort needed for it to just work". That's one of the major issues of webOS, Android, and iOS: most of the stuff is a self-contained "app", almost nothing's integrated.

EDIT: look at all the crap standalone messengers on Android, and the story behind the (still!) missing ICQ/MSN/XMPP plugins for Synergy Messaging on webOS - nobody took the effort to write a proper plugin for now (webOS 3.x), but some people just want to sell their crappy standalone stuff instead. Same with GStreamer vs. standalone mediaplayers w/ built-in codecs or proprietary streaming solutions there.

You too are missing my point. I'm not opposing his opinion of having the thing integrated into maemo. I have a N900 and one of the top features I love about it is the way its messaging work. What I'm saying here is that reverse engineering a protocol is not a simple task. You should focus at least on writing a code that actually works first (at least logs in, and sends and receives messages correctly), which wouldn't necessarily have a GUI, and then worry about how the user frontend would look like (whether it's gonna be a standalone app or integrated into maemo's messaging system). You can't just focus on the user-end when the underlying engine itself does not exist! First things first guys, this shouldn't be hard to understand if you're a developer.

My message was more aimed at the people that are still trying to figure the protocol, like a proof of concept client. Seems like teamer figured out everything, so that's great having to depend just on him For the people working on the protocol I have got a Nokia C3, I've just installed the latest S40 whatsapp client and will try to capture something.

I personally believe teamer hasn't made any progress further than registration (which was actually posted by colacheater in first place). And I think many devs here would agree.

The protocol appears to be just bog standard XMPP (I'm currently reading the XMPP Definitive Guide book and the RFC's).

Registration and sign-in is of course specific to whatsapp, as can be expected.

Anyway for those that don't know what the XMPP protocol is, here is a 5min intro to it and assumes you have some dev knowledge:

An XMPP is just a long lived TCP session (a long TCP session).
XMPP is a profile of the XML standard.

When you start an XMPP session or chat you send a single open session tag. This means the session is open and you can communicate.

Each time you send a message you send a STANZA - this is basically just a block of XML. Thus your session will never be well formed XML until you send the closing session tag.

This is the basic core of XMPP. Additional stuff like multi user chat, sending files and realtime video/voice or streaming etc. are enabled by extensions - which you can query the server to check what it supports.

Check the different RFC's for more detailed info - perhaps it will help some of the guys working on it here.

Post any Q's you have here - I'm not an expert on XMPP but I am reading the book, so will try to answer. Feel free to add anything you think adds value to my post or anything you think is not quite right.

the question is: since whatsapp port might be developed as a telepathy plugin does this mean that it can be used on the N9 MeeGo-Harmattan OS?

I certainly hope so as I'm using a n9 now!

Just because of an PN I got I want to give, independent from the rest of the dev-talk here, some thoughts on security with the to-develop WhatsApp client.

Everyone of us who had a little bit intenser look at WhatsApp should have noticed some security leaks I think. Maybe even more than the one you find scattered on the web.
We may should spend some thoughts about how we could try to prevent everybody using these security issues to fake others accounts, log into existing accounts and so on.
We shouldn't be that naiv to think that WhatsApp would be more secure when we don't make these public but what I mean is just preventing that tons of script-kiddies are playing around feeling cool that they can hack WhatsApp. As said: everybody really willing to find the security isses will find them. I also think that if we wouldn't totaly crash the "security" system WhatsApp has we may be more toleranted by WhatsApp regarding the inofficial client.

Security always is a very difficult theme to talk about, especally how to proceed with found issues.
For my part there are two main reasons why I would try to kepp them "secret" in our case: We actually are using some of them to make the inofficial client work and the said prevention of script kiddies and with that maybe the friendliness of WhatsApp.
Issues we don't need to make the client work and also are minor issues we may can report to WhatsApp independent from our client-work without link to the threads here etc. (but as the history of found issues shows they seem to don't really care about them)

The difficulty I now see is that if we would write in public what these issues exactly are so that the one working on the client(s) can consider them in their clients we would make it in ways needless. But writing them just a few persons always excludes the other developers.
How the single finder of the security issues is handling it is his decision I would say but at least I woudn't write an "How-To hack WhatsApp"

For my part the issue I think everybody is aware of and is a minor one we should consider is the registration of new accounts.
With the known way we can fake accounts with numbers we don't own.
The idea of letting the user choose if an automaticly generated password should be used during the registration or if an own password should be used (make an md5 of the password and it shouldn't stand out during registration) is an option I would support.
But everything else that connects the account to an specific phone should be left by the default as the WhatsApp client is also doing it.
This way should be okay for everybody wanting an WhatsApp client for the N900/N9 I think (and it would be the easiest to use for an non-developer). There aren't more options you have to change. Maybe an option of hiding or sending the MCC/MNC would be discussible but I would just use 0's as default.


I hope that every developer currently working on WhatsApp agrees with my position, at least the main ideas.
If not it would be nice to have a small discussion about security here regarding how far we can offer options to the client without threatening the abuse by some people feeling cool because they can do things they shouldn't do...