I mean, what gives with that? 

It just isn't very high on the our priority list. In short, it's a lot of work for very little gain.

The long version is this. Changing to a universal binary is very simple if that was all we had to do. However there are a number of other things that will be affected, including our auto updating feature that we have in the Notifier. This is the tricky part that will take a lot of time to get done correctly. It will be completed one day, just not important enough to get done right now.

FYI: I have a universal binary version of the Notifier running and it doesn't noticeably perform any better than the non-universal version. Hardly surprising since the Notifier was never slow in the first place. I am assuming you want the universal binary for performance benefits.

Some users have told me about issues with having long running Rosetta applications (running non-universal binaries on an intel based mac), I'll have a look into this and see if it justifies releasing a universal binary sooner.

I don't understand your reasoning on not releasing a universal binary if you have one already compiled. It's not about performance, it's about stability.

Running Rosetta means a whole load of code loaded into the system, slower startup (which impacts when the notifier is installed as a Start at Login item, because as well as loading the notifier it has to redo the Rosetta translation) and.... 

Well, you *already* as you have said, have a universal binary of it compiled up. Now, I understand the desire to make it part of the next version released, but honestly, you were saying this back in 2006. 

And I speak as someone who's written notifiers for other online services.

Oh, and I don't understand how the autoupdating behavior between a PPC app and a Universal app would change, or are you saying that the current notifier has an older behavior which you want to engineer out, and you want to use the next release as a "new updating and Universal" release?

Minster just got back from his holiday to the great wall of China, I think the real reason is he has forgotten to code.

Nice holiday!

Sorry codepope I'm still stuck on windoze, so there is nothing in it for me.

codepope: As I said, compiling a universal binary is the easy part however it introduces other problems which needs to be addressed. One of them is the auto updating feature on the Notifier. It just isn't so straight forward when it comes to a universal binary.  I can't really explain any further without revealing some of the inner workings of the Notifier which I am reluctant to do.

I can understand your concern about the stability of Rosetta and I can see similar websites reporting the same thing with their application. However, with my own personal experiences, and this comes from someone who has a major role in the development and has run the Notifier for far longer than anyone else, I can honestly say that our Notifier runs just fine under Rosetta emulation. Not optimal I agree but satisfactory for now. Until I have more hard evidence that Rosetta is causing undesirable performance or stability issues in particular with our Notifier, then developing a fully supported universal binary is very low on our priority lists. This is the reason why you have not yet seen a universal binary, it's relative importance to other items on our to-do list is simply very low. It will be done one day, just not anytime soon.


And I speak as someone who's written notifiers for other online services.

And I'm sure you will appreciate that our Notifier has different features and requirements to other notifiers. It's hardly fair to compare our product with others.

Well, yes, I'd guessed that the Notifier was a bit different from most notifiers what with the Sqlite3 database and persistent socket connection doing what appears to be two way XML communications. (Hey, so I'm curious... I looked in .tangler.... ) And yes I did notice your handshaking has MAC-PPC hardwired into it's name so I am guessing thats where your issue lies with changing the notification.

It feels more like a framework for a locally replicated view than a notifier, except you aren't retaining the messages... I guess thats an option for you chaps in the future... if the notifiier API was open I'd be off and writing a Java desktop client for it; it looks pretty capable.

While I'm at it can I just report a bug; the users password isn't hashed in the local db so it's readable. 

As you can see, the Notifier isn't as secure as we like it to be either. SSL, password encryption are already on the to do list and are ranked pretty high, at least more so than the universal binary issue

Notifier API is definitely possible, unsure if it will ever happen though. Hasn't been discussed internally here yet.

An API is definitely on the cards, which would probably be offered as HTTP REST style architecture. Not so good for pushing notifications out.

Minhster, you seem to be talking about the new version of the notifier and the universal version of the notifier as two separate things. Surely it would make sense to make the next version universal otherwise you just seem to be making excess work for yourself...

@Craig, well you could consider a long lived COMET style connection in the REST mix; anything is better than constantly polling. 

@Minhster, SSL would get past sniffing attacks, but hashing the password is quite a cheap (in terms of complexity) fix, just tweak the login process to accept an unhashed or hashed password (if you don't have space in the protocol, cheat... say a hashed password is just one that starts with ###). If you aren't hashed in your internal db, then all you need to do is take the internal password, hash it and compare it with the password just sent to you less the '###'. Then clients can just migrate to using hashed passwords.

SSL is a different kettle of frogs, and it's more of a scale issue at your end (because heck, if you SSL the notifier, you are going to have to have SSL on the web interface for it to make sense)....

 But whichever, I really think it should be a fixture of your next notifier release for Mac that it is universal. I mean, you don't want to be associated with say Microsoft who are also taking ages to go universal do you

nomine: I'm not quite understanding what you mean? We've had plenty of new notifier versions that has not been universal binaries...

codepope: Sorry but this is where I leave this discussion. Universal binary will be available for download when it is ready and not a moment later

So you've just been putting off making a universal version for months on end. The work still needs to be done, why not just get it out of the way as part of building the next version.

hehe you are assuming the only thing I work on here at Tangler is the Notifier....

No i'm not

How about implementing an API for this functionality? It can hardly be more work than writing your own and I am sure there is plenty o people who would quickly outdo your current notifier. Just a thought

Alex: As mentioned above, API is possible, just never really discussed here internally yet. So I'm not sure when it will be done.

oops.. must have missed that part.