“I have a vision of an application which provides me with an idea of the sort of features I would like Syllable to have. Its time I wrote it down so that you can all see where I’m going with Syllable. If I compare my fantasy application with the current abilities of Syllable, then it gives me a set of features which we still need to implement, so it sort of serves as a roadmap and a high level design all in one.” Read the article at Syllable’s web site.
Damn. What he describes in this article are the services in GNUstep/Cocoa… look at http://www.stepwise.com/Articles/Technical/Services.html for a good description. Basically, every program could vend some services (send by mail, send to ftp, etc.) and others programs could access for free to theses services (they not even knows what services are there, it automatically appears in the menu services). A great idea.
Doesn’t MS Office already do this? Whats so great about it?
Martin Luther King revised? (“I have a vision…”)
Microsoft uses various systems such as COM (Component Object Model) to provide the sort of functionality that Office has. On the surface, yes it is similiar. In reality, the model I describe for Syllable is higher level, and relies on Message passing. It is, in fact, much more like Rexx (Which was used in OS/2 and on the Amiga as ARexx). It really is quite an obvious and easy way to get application integration with the client & application server model we have.
I’d like to add that it has been interesting to see the sort of ideas that the article has already sparked off, and it seems that it has helped to get people thinking along the same lines, too.
Well, I guess shaping the future these days deemed much easier than writing hundred lines of kernel code.
I’ll just stick with stupid people like Gates. No brains to “shape the future” but no agenda either.
Get real people
I agree. But I think the MS model is better. With COM you can add office functionality to your own programs rather than having to call an external program to do this on your behalf. For example, I can add an IE’s HTML translator into a program of my own.
Yes, but why doing Message Passing ? OpenStep does exactly what you describe, by using services, which are based on the pasteboard. imho it’s cleaner and more powerful than fixing potential messages (services works on datas, images, strings, etc. (everything which is possible to put in the pasteboard)).
In working with the type of datas, you are free to send it to whatever service which accept the same type of datas.
it has been interesting to see the sort of ideas that the article has already sparked off, and it seems that it has helped to get people thinking along the same lines, too.
Well services are one of the most interessting thing of OpenStep, yet a simple one. I could select some text here, and send it by mail just by selecting the service provided by GNUMail … Terminal.app provides a method to declare as services shell commands, etc. This is really a great way of make apps collaborating while remaining simple.
Dear Vanders,
It seems too many developers are so knee-deep in the tools and tricks of programming that the actual user experience is greatly ignored. Thank you Vanders for your genuine end-user desires meshed well with your technical knowledge of what is possible. The world needs more well-rounded developers/users like you.
Just a short correction (if I’m right):
“Glasseye can take advantage of libnet very easily. All it has to do is create a new FTPStream class,”
should instead state:
“All it has to do is create a new FTPStream instance”.
PS I really would like to see this when it gets done.
Yes, but why doing Message Passing?
Simply because it is a simple extension to the model that we currently have. The application server communicates events to the the applications via. Messages. It is a simple task to extend this form of IPC to inter application messaging.
OpenStep does exactly what you describe, by using services, which are based on the pasteboard. imho it’s cleaner and more powerful than fixing potential messages (services works on datas, images, strings, etc. (everything which is possible to put in the pasteboard)). In working with the type of datas, you are free to send it to whatever service which accept the same type of datas.
Messages can also encapsulate many data types, including raw binary if required. Syllable will use the Registrar to ensure that the application can handle the data that it is sending, and any registered application should accept the data it is sent. In practice, it would not appear to be much different to OpenStep (Although I admit to not knowing anything about OpenStep services other than what you have described here, and some snipets from the link you provided)
“All it has to do is create a new FTPStream instance”
Quite right. Thanks for the correction
“Simply because it is a simple extension to the model that we currently have. The application server communicates events to the the applications via. Messages.”
“Syllable will use the Registrar to ensure that the application can handle the data that it is sending…”
Application server… Messages… Registrar…
This reminds me of some other OS )
Looking at their screen shots….
FileSystem Type: fatfs
Size of Partition: 4.2 Gb
Percent Used: 724.8 Mb
Percent Available: 3.5 Gb
#1 – 4.2 Gigabit drive? Wow. Didn’t know they made those. I’m sure they ment Gigabyte (GB).
#2 – 724.8 Mb percent used? Mb must be a new “percent” demarkation. 724.8 percent used, though? Their FAT table must be screwed up.
#3 – Megabyte is MB not Mb.
Who are these people? Highschool drop-outs?
Also, someone should clue the author in about existing technologies – he’s re-inventing the wheel.
Old screen shot, that has been fixed.
Yes, it may be re-inventing the wheel, but do you have the source, time and effort to help us bring this to Syllable. Yes this vision is close to BeOS and COM, and … but most of these solutions are closed source. Also this is a hobby OS, we maybe targeting the home market, but we aren’t trying to take over the world, especially with Syllable not being close to a 1.0 yet.
Is it just me or is there a general problem in displaying Syllable’s website? My browser (Netspace 4.5 on Win95) only shows the title and a horizontal icon menu bar, but no article. I’ve tried to get the FAQ but with the same result. Any idea?
I am the lead developer of the desktop and I will assure you that those errors are fixed. I made a couple mistakes, because I am basically writing the whole desktop myself. I do a lot of work for Syllable and I am human so I do make mistakes As for being a Highschool drop-out, yeah right! I am attending California Univ. of PA and I have a 3.75 GPA, so no I am not a Highschool drop-out!
You obnoxious tone suggests that ironically you are probably still at high school.
I had a similar idea a few years ago. I needed a supervisor for the various services I’d be needing.
GNATS turned out to have all the qualities I needed.
So I’d say, take a good look at the various bug-tracking systems and if you can, at the various transaction monitors available, beofre setting code to comp0uter.
I really do not agree reinvent the wheel again. I prefer to use re-produce or re-create. I don’t think a wheel produced in 1900 is the same as those produced in 2002.
Thx.
can’t anybody just make a useable os before the grande visions?
I could post a comment here and defend Syllable against snide comments such as those by John Galt, but then arguing wether Megabyte is Mb or MB is just too pety to bother (Hang on, shouldn’t we be using SI units now anyway?)
Anyway.
I never wrote that peice as a general “Hey, lets take over the world! We rule!” article. I wrote it so that other Syllable developers can see where I would like Syllable to go. If you read it and find it interesting, then that’s great. If you read it and think it is silly then just ignore it until we are done. Then you can try it for yourself.
To a certain degree, yes it is reinventing the wheel. It isn’t really a component technology like COM, though, so we’re not reinventing that. Its more like Rexx for 2003, with embelishments that make it look like a component model. If you look at the client-server model that Syllable uses, then it makes far more sense.
The whole point of it is to make Syllable useable. A lot of the lower level parts are done, or in development already (E.g. the kernel, GUI, supporting servers etc.) so we’re now looking at what we can do with all this stuff we have. We’re trying to create a design. If you don’t like the design, then please join in the discussions and help us to make the design better.
this sounds like a great idea that has been kicking about for a bit, but doesnt appear to be widely used. it’ll be great when you make it work!
1) can the pasteboard/clipboard functionality be subsumed into this model? when someone selects code you send a message “new selection: foo” and if the standard clipper can’t handle it (maybe its an image) then the system server starts up an enhanced clipper?
2) new applications should register their abilities in a standardised, yet extensible way. when i download the latest “foobar” application which allows me to “foobar” my documents i want that functionality available, even if noone thought about it and there isnt a #define anywhere for it.
3) when applications install and register, if they offer functionality which already exists then the user should choose the default. there is nothing the application can do to make itself the default.
No you cannot.
For that developers need to be able to program.
That was left to Atheos guy (hoping he’ll resume code then reap his code off), they are visionaries.
If you need high level expertise on kernel development or politics, get in touch with these people, but do not bother asking them for stupid functionality.
:-()
I noticed the same error as you. But then I noticed, hey! They must’ve fixed it:
http://syllable.sourceforge.net/modules.php?op=modload&name=My_eGal…
Well who knows, it looks promising, I won’t use it for awhile. Hopefully by then there will be some other promising OS’s to try out as well.
Interesting.
About the standardization stuff, it would be easier if the design allows for a translator type of implementation. This is very similar to the model following BTranslator ( http://www.tycomsystems.com/beos/BeBook/The%20Translation%2… ), though. Just specify the interface and let the vendors plug in their own implementation.
Good luck on your project and happy holidays to all.
For that developers need to be able to program.
That was left to Atheos guy (hoping he’ll resume code then reap his code off), they are visionaries.
If you need high level expertise on kernel development or politics, get in touch with these people, but do not bother asking them for stupid functionality.
:-()
Maybe you should actually look at the differences between AtheOS 0.3.7 (The last release) and Syllable before you shoot your mouth of, kid.
If you would like to astound us all with the millions of lines of code you’ve written in the past six months, please do. Or would that be the reason you havn’t helped out, because writing code is harder than making snide comments on OSNews?
Thought so.
I did like the kid part.
It makes me feel young.
I do program for living. As far as help, I am helping *BSD efforts. It seems far more rewarding for number of reasons:
1) useful today
2) people who work on it know what they are doing
3) available under liberal license which allows me to use it at work without having to distribute code, unless I want to
I think person who could write millions of lines of kernel code in half a year which is roughly 15,000 (that is if we re talking about two millions) lines per day working overtime should be at least as smart and old/experienced as you are.
I am not even pushing for such an impossible goal.
Could I get some candy money of you, pro?
BTW, how old are you 19?
I had prepared a long and detailed rebutle of your post, Buddy Love. Then I realised that it doesn’t matter what I post, because you’re being a petty troll. (A licence flame? Oh please)
If you’re really interested in Syllable and our abilities, come look again in six months time. If you’re not interested, just close your eyes and stick your fingers in your ears.
1) The BSD licence is like working for Bill Gates without pay.
2) BSD is useless to anybody who isn’t a geek
3) The person who made AtheOS probably knows more about what he was doing than just about anybody other than the likes of Linus Torvalds and Andrew Taunbumn (sp?)
4) Syllable is for the desktop, BSD is for the command console.