If there was a single application making someone use MacOSX full time, that application would be Watson. This application is possibly the most awarded in the MacOSX world, and it is the best idea for an application that I saw the last few years. Despite the idea being dead simple, and even seem redundant at first, Watson is an excellent, original and handy tool. I would like to initiate an open source project for a multi-platform clone of Watson (Karelia say that they have no plans for any ports – with only a remote possibility for a Windows port). I will start by putting $100 USD on the table while readers are most welcome to co-sponsor the project as well. What we really need though, is willing and capable C++ developers.If you are interested to develop the app, please send us an email or reply into the forum about your intentions.
If you would like to contribute money for project, please add as much money you are willing to by sending us an email and telling us about it.
The reason why the Qt API was picked is that it is truly multi-platform (providing better native support under Windows and Mac than GTK+ or even wxWindows), and it is able to run on many embedded platforms. Being very portable is a key part of this project. (QNX, OS/2 and BeOS users should also be able to run the app through Qt and an X server; while not as good as a native port, it’s better than nothing).
The developer(s) can pick whatever name they want for the application. Some ideas: WhatsOn (as opposed to ‘Watson’ – but this is probably not a good idea, plus there are no domain names available for it) or WebLeech (freaky, but accurate).
Rules:
1. The application must be written in Qt 3.x and C++. It should be as OS-independent as possible; OS-dependant code should be clearly seperated from the core of the program. For example, you can use the KDE libs only in places where they are really needed.
2. The application must be also developed, or at least written in such a way that it will be easily ported (possibly by third parties at a later time) to Windows, MacOSX, Unix/KDE and plain framebuffer Qt (for the embedded devices, eg. Zaurus) platforms. In short, wherever Qt is available.
3. The application must allow the use of modules or plugins, so other programmers can write their own services. For example, another third party could develop a module that is able to leech the French telephone directory and then easily install it by either placing it in a subdirectory of the app, or installing it automatically through the web.
4. The application must be under an open source license: BSD, MIT/X, GPL, LGPL, Public Domain, Artistic License, OSI Approved, Mozilla, zlib/png, Apache. Pick one, I don’t care which one. However, please note that if you prefer GPL/LGPL, given the multi-platform nature of this project, you must be aware of the Visual Studio.NET issue. So please decide wisely.
5. The money collected will be given to the developer or to the leader of a team of developers through Paypal after they have completed at least 3 modules and the plugin engine is ready to be used by other third parties to write additional modules. The modules should be full-featured, fully-working and their subject must be 3 of these.
6. OSNews will be able to provide web space to the project, at least for its first steps, or you could use SourceForge or an equivelant freely available service. If you are a single developer, you are free to develop it locally on your PC, but it would be nice if you could post a source snapshot each week, until the app is ready for release.
7. Only genuinely interested developers please proceed. You may use our forums to find partners and consist a team if you do not want to work alone.
8. The application is to follow the Linux versioning model. 0.x for the alpha and beta versions, and then, 1.x versions for the stable ones.
9. OSNews will promote the project by freely providing ad space for 10,000 ad views per month, for 3 months. News items on the development effort will be posted every now and then, to further strengthen the project’s publicity.
10. The interface must be created fully with C++ and Qt. No HTML/DHTML interfaces via the Mozilla/IE/Konqueror/etc engines please.
…open source software does not pay bills or buy food.
This is why we, the users and OSNews, are trying to sponsor the project by putting money on the developer(s)’ PayPal or other (bank) account.
Damn, this idea is so freaked out, just go and fucking buy a Mac and support the developers instead of making their lives more uncomfortable and developing an opportunent.
BTW: What kind of journalism is this, putting your own opinion into your articles?!
> Damn, this idea is so freaked out, just go and fucking buy a Mac
Not everyone has $1100+ to spend for a new computer, when they already have a computer. This proposal was never meant to make the lives of Karelia difficult, but merely to provide an opportunity for PC users to have an equivelant application! If Karelia had ports for Linux/Unix and Windows, there wouldn’t have been such an project initiation today.
> BTW: What kind of journalism is this, putting your own opinion into your articles?!
An editorial/feature maybe?
And, what opinion??
The $100 USD will go off my wallet, not OSNews’ (I do not manage the financials at OSNews, neither OSNews is mine). However, *I* initiated this project (with the support of the rest of the OSNews members, as I learned earlier today
so if I want to say that I like portabello mushrooms more than the wild ones, I will say so.
And, watch your language.
sounds like a worthy project although I’m not entirely clear on what exactly watson does.
I unfortunately dont have the c++ skills to help out but I hope someone is interested enough to do it.
Damn, this idea is so freaked out, just go and fucking buy a Mac and support the developers instead of making their lives more uncomfortable and developing an opportunent.
I’m not sure that would be many Linux or other Qt enabled OS would gladly buy a Mac and use OS X just because of Watson. The developers doesn’t want to port it to other platforms, so why not just have a open source version?
BTW: What kind of journalism is this, putting your own opinion into your articles?!
It’s an opinion, not a news story. If you actually read the entire proposal, you would realize the first word in the title is “Proposal” Go get a brain.
—–
4. The application must be under an open source license: BSD, MIT/X, GPL, LGPL, Public Domain, Artistic License, OSI Approved, Mozilla, zlib/png, Apache. Pick one, I don’t care which one. However, please note that if you prefer GPL/LGPL, given the multi-platform nature of this project, you must be aware of the Visual Studio.NET issue. So please decide wisely.
Except maybe to recompile the program under Windows, I’m sure no developer would use VS.NET for QT Development in the first place.
But anyway, if I was to pick on, I would pick the BSD License. No, not that I’m anti-GPL; here’s my reasons
– Watson depends a lot of their own servers. If people who just take the code and build their own web services app without releasing the source, it would be more widely used. A lot of servers for sure wouldn’t want the source code of their version out because sensitive stuff like server protocol are in there and freeloaders can easily use the server without payment.
– It is a lot less complicated. And you could use VS.NET anyway.
– You might help a company earn a buck 🙂
(yeah, yeah, you caught me, I only have one proper reason).
8. The application is to follow the Linux versioning model. 0.x for the alpha and beta versions, and then, 1.x versions for the stable ones.
I think the versioning model should be the developer’s decission. After all, they are the one setting the versions, there is no use using a versioning model that they don’t understand.
9. OSNews will promote the project by freely providing ad space for 10,000 ad views per month, for 3 months. News items on the development effort will be posted every now and then, to further strengthen the project’s publicity.
Whoa, that is worth a lot!
But would these ads start when the app reach 1.0 or would you start promoting an unstable app?
The developer(s) can pick whatever name they want for the application. Some ideas: WhatsOn (as opposed to ‘Watson’ – but this is probably not a good idea, plus there are no domain names available for it) or WebLeech (freaky, but accurate).
*Warning: Very shallow* How about RajanR in honour of the great me?
Anyway, I got a good idea: Hallazgo; spanish for find. I got a logo idea, and would try to draw it out.
Thank you for your comments RajanR.
> But would these ads start when the app reach 1.0 or would you start promoting an unstable app?
Whatever the developer(s) want!
…open source software does not pay bills or buy food.
There is money to be made: servers. Yes, OSNews is providing servers for development; but not for services like plane schedules and weather and so on. So you could put the content into the project and charge for it. And you could pay bills and buy food (though you could start farming) with the money you earn. And because there would always be competition, you would always be on your toes and providing the best fastest service.
Heck, i’m giving away an entire business model 🙁
Poor thing I have no money to start any business.
of what watson does…tho i read the PROPOSAL and the website… but hey, sounds cool!
Just look the screenshots ( http://www.karelia.com/watson/plugins/toolbar.html ). It is an app that accepts plugins which are able to “talk” to certain websites and retrieve results for a query. For example, it can leech on a tvguide.com site, and reply for you where and when Star Trek will be played on.
Or, you can try to find tickets for a movie.
Or, find a person’s telephone number.
In short, it takes some services from the web, that are available through a web browser, but simplifies them in a nice GUI and puts them all together in the same app!
As I said, it may sound redundant to do so, but believe me, after you spend 2 minutes with Watson, it becomes a must-have!
Watson looks like a handy tool, but prone to breakage as the underlying sites/interfaces change. In some ways it is the “anti-evolution” tool in that it forces web providers (sites & services) to keep the same. Or at least be more honest about keeping their old interfaces available 😉
It would be fun to create a Windows version of this tool that leveraged all of the cool UI stuff you can do in WinXP. I’m not too interested in doing a cross-platform version. Linux has a completely different UI and a lot of this sort of application is actually not on the data gathering/processing side, but on the UI display side of things.
Good luck with the project.
#m
… This sounds kind of fun.
I’ll speak to a few local friends about this. he he
james
> Watson looks like a handy tool, but prone to breakage as the underlying sites/interfaces change.
Agreed. It will need constant attention, on the other hand, these big sites do not change their web engines…
Also, this is where real Web Services can greatly help. By using the SunONE, or the IBM WebSphere or the .NET tools to create these web sites, the app wouldn’t break because of the XML help!
Good to see users putting forward good ideas and sponsoring them
I’ll be watching to see what happens here as most of the so called business models I have seen for open source I am still skeptical of.
It will be interesting to see what karelia makes of people sponsoring a opensrc app in direct oposition to theirs.
——————————————-
http://witme.sourceforge.net/libferris.web/
I looked into the project a bit more while I’m waiting for some files to download —
The bugaboo of good cross-platform libraries is still price. Just a single developer license of Qt for three platforms (Linux/Windows/Mac) including support is $4000.
BTW, I think it’s admirable of you to put your ideas and desires on the table. It is the real American way 😉 Competition is a good thing!
Best,
Michael
HI!
This App certainly looks nice and wortk the effort of getting it across platforms, but what is even more interesting to me is concept of user sponnsored Open Source software development!
After the lost of Cosource.com that had this service of writting down your brainstorming results into project you wanna see developted and putting the money in there were no other attempts. This is really bad as a lot of GNU/GPL software could get sponsored development if micro sponsoring would be more simple and structured.
My opinion is that Cosource didn’t make it as this was the only service thay did… if Sourceforge was to take on this it would have been a success for sure!
Does anyone know of services which make micro sponsoring more easy to do? Or has ideas how this could be inplemented?
Best,
Zeljko!
Uh what a lot of cr*p!
This project is only for the Mac wannabes.
Leave me out of it.
From Karelia’s FAQ:
Q. How does Watson relate to Apple’s Sherlock?
A. Watson is similar to Sherlock in general functionality, in that it puts a Macintosh façade over web sites/services. Otherwise, there is no relation between the two software products.
The bugaboo of good cross-platform libraries is still price. Just a single developer license of Qt for three platforms (Linux/Windows/Mac) including support is $4000.
That is for propreitary/ commercial software development. In case you didn’t read the article, it is open source.
Uh what a lot of cr*p!
This project is only for the Mac wannabes.
Leave me out of it.
Then I guess you don’t use a lot of applications, like GIMP, or KDE because they are a <application>-wannabe.
In fact, you are probably not using any Linux, FreeBSD, OpenBSD or NetBSD because they are a UNIX-wannabe etc.
This proposed project is to deliver the same features/functions as Watson (which is very different from Sherlock) but at a possibly different UI, layout and presentation. Heck, this may be even better than Watson.
Why can’t programmers separate the core code from the interface, so that they don’t have to lock onto a single graphic toolkit? The Qt vs. Gtk thing is getting tiresome. Whether I choose GNOME or KDE I loose, because only half of the best apps run correctly integrated on my desktop.
Please practice general programming, and give users more options.
> That is for propreitary/ commercial software development..
The MacOSX/Windows versions of Qt indeed cost money, no matter the nature of the application (open source or not). Qt is free only for Unix, and only if the app is open source.
But there are developers out there that they do own the software needed, and they may want to port it at a later date to these platforms, after the Unix version is done. This is why I insist in OS-independent code, so the easier something is ported, the more chances they are these developers will give it a whirl and recompile it for OSX or Windows.
Sheesh, so I learned Windows and then branched out into Mac and Unix 😉 For me, the Windows version would be fun to do first considering the vast array of cool UI capabilities on the Windows platform. And to start doing a Windows version of Watson with Qt, there is no default free path.
From having worked on a large GUI intensive cross-platform C/C++ application that ran on Windows (16 and 32 bit), MacOS, and 25 kinds of Unix, let me only offer my experience that a bit of cross-platform design/prototype work upfront saves major grief later.
From my experience in building/porting cross-platform C/C++ apps, I try to design the most feature-rich edition of the application first. It is easier to subtract out features vs. build them in when the core engine doesn’t support them.
So, yes, for Linux (Unix/X11), the free edition of Qt is available. It is a place to start where the entry costs are low. The important thing is actually starting. Unfortunately for me, I don’t have a Linux dev box setup. Perhaps this week I will dust off my old dual P3-450 and get Gentoo running on it.
For now, goodnight!
Thinking about this, I can’t see why you wouldn’t use Java!
Its everywhere you want to be (except Be), including the PDAs & phones from the other article.
The net is full of Java folks who use it for server stuff & services etc.
No pay license needed, everbody knows it, & its 2x productive over C++. Lots of tools for it even still on Windows.
I have trashed Java on occasion for big GUI graphics apps, but from what I can see, this is almost entirely non gui, the modest screen shots I saw would be easy enough for Java Swing or AWT.
Note that Java & Cocoa are considered to be near cousins in the OSX world, I would choose Java over Cocoa just to be independant of Mac even if Apple folks want Cocoa.
Remember those plugins, if you want to use the plugins already done, I am sure it would be easier to convert them from Cocoa to Java if those developers wanted more platform access. It may be possible to work with those plug as is.
You could do initial development under OSX & clone the App in front of your eyes. As it comes along, retest elsewhere.
Apple’s Java is supposed to be 1st rate now, its one of the things that might draw me back to Mac dev. For Windows, you could compile it as a native .exe for speed with VCafe4, use my license if you want.
Name Sherlock x, Watson x,
how about Moriarty?, well he did become sentient in the STNG & went off exploring his own little universe with posh babe!
JJ, good points about using Java. That would definilty be the way to go for a project like this. What ever you lose in performace, you gain speed of development and porting. Also, has anyone here seen the Swing2 library from Sun? It makes a very nice looking GUI, something could seperate it from the rest of the half-done apps floating around for Linux/Unix.
I’d have to say that your idea is very good. To me, it seems that an app like Watson would be eaiser for me than surfing some webpages (There are just bad UIs all over the place!).
And JJ, the episode your talking about is entitled “Ship in a Bottle”. You earn major cool-points just for metioning ST:TNG
. I vote that the name should be Moriarty!
Whatever might be said for Java on the server, it is a very slow and cumbersome environment for running an application on a client computer. NeXT had much of a good thing with Objective-C 😉 Code that runs on the processor itself is a good thing.
IBM had to rewrite the base windowing system to get reasonable performance for the Eclipse IDE. A software developer will take the 33+ MEG hit to run a fancy IDE. The casual user typically will not. It takes a Java expert to write a high-performance client application. That’s probably more rare than a decent C++ developer who can use the Qt framework.
As for the portability and prototyping, I agree, Java would be handy. The tools are top notch, especially JBuilder6. And the runtime cost is right. And there are no funky licensing terms as you get with the “free” version of Qt.
Anyhow, the more approaches the merrier. Perhaps the Java team can race the C++ team?
As for Java being 2x the productivity of C++… when I went to XP Immersion and was part of the sole C++ team, we beat every single Java team. In fact, we were the only team to finish all the assignments. It is a BIG MYTH that Java is any more productive than C++.
The non-PC reality is that many programmers become Java programmers because they aren’t talented enough to write software using C++. In all my development experience with Java and C++, I’ve found the opposite to be true. Java leads to poor productivity and poor code quality more often than not.
Anyhow, not to start a Java vs. C++ conflagration. Let’s all get started on what we want to do. I’m in the C++/Windows camp for the most part. Perhaps C++/Linux if I get my devbox running this week.
Adios.
Why java? I’m sure Eugenia doesn’t want to wait a few hours for it to load and grab files. ;>
low exaggerated punch i know, but it was open. hehe
Why not do it in delphi/kylix? You’ll at least get a linux and windows port out of it. Let the guys making the original have an opportunity to get paid. I know it misses alot of other os’s, but freebsd has a compatibility layer right?
Hey, I used to work for Borland. So, I’m down with Kylix. And if what I hear is true, there may be a Mac version of Kylix in the works, giving you all three platforms with one code base. And you’ll get more and more interop between C++ and Kylix as time goes on.
Not to forget… Kylix generates actual code, not byte code. And is designed and built by some very cool people who will only make it better and better.
Good point. Much better option than the 4 letter word.
#m
Write the core in Object Pascal and you could port it to Solaris, BeOS and QNX as well as *BSD and, er, OS/2 and DOS. There are ports of Free Pascal for each of these platforms (Linux/Windows is covered by Borland, though Free Pascal is also an option).
There is no advantage what so ever coding it in C++. Eugenia is a big fan of C++, so I gues she is a little biased.
The thing that worries me about this project is the line from the Watson FAQ:
> Q. Will Watson be available on Mac OS 9? On Windows?
> A. We love Mac OS X so much, we’re planning to devote all
> of our time to it! A third-party port of Watson to
> Windows is a possibility.
Why don’t you all save yourself the hassle, pool up some cash, buy the right to port the software and give the poor developers their dues. If I were Karelia, I would have copyrighted the concept/technology/IP, and would sue pants off of anyone who tried to undermine my position. You will litterally be stealing the money from them.
If this goes ahead, and uses anything less than the GPL, people could proffit greatly from it, and that’s just plain wrong. It’s a sad world when the only good ideas are those we steal from legitimate business for our own gain.
You could always just bookmark the respective sites. I really don’t want to see Watson developed on other platforms (really just not Windows – I guess a *NIX port would be OK) because if it gets to popular, the corresponding sites that provide the services will get pissed that no one is seeing their advertisements. Then, either Watson would get crippled by the sites or be forced to include ads in the interface. On the other hand, if you look at screenshots for the new Sherlock in Apple’s next release of OS X, it seems that Watson has been incorporated into it – so it may end up becomming a legit partner with these sites anyway.
As I said, atleast Windows versions of Java can be compiled to native .exe so speed of simple gui apps should be close to C++ apps, thats my experience (assuming not too much new()).
For full blown Gimp style apps, forget it. Clearly this app has a fairly small front end gui that I know will be fine. The backend stuff, I don’t know that stuff, but surely most of the delays are going to be waiting for responses (most users are still 56k).
Consider this, isn’t this the type of Java app that will end up on PDAs & cell phone, if ever a road warrior needed Watson capabilities, how would those guys get it into such an underpowered device. x86 is going to run at least 5-100x faster than any PDA so I don’t think speed is going to be much of an issue. If Java can’t do this on a 1G x86, it sure won’t be much use on a PDA/phone. Embedded Java maybe another story. I don’t think rookies usually write the important stuff.
I was under the impression that Java was vastly more filled out (ie bloated perhaps)in the classes dept than Qt.
Where does the 33Meg come from, the developers env doesn’t count. A compiled Java .exe ships as is (I am not sure about dlls).
“It is a BIG MYTH that Java is any more productive than C++.
The non-PC reality is that many programmers become Java programmers because they aren’t talented ”
Whoa, well you said it. I was using the 2x for experienced developers comparing C++ v Java/C# for same guy. A rookie Java only guy may well produce crap, maybe, I don’t know any rookies.
Myself I will be sticking to C++/asm mostly as well, and playing with Java to see if I can save some legwork.
Good luck with the project!!
If you were to develop this software for Windows using the non-commercial Qt for Windows, and you were to take the $100.00, wouldn’t you be violating Troll Tech’s license? They require that you develop non-commercial software in a non-commercial setting. Further, they say:
“A non-commercial setting means that you must not use the package in the course of your employment or whilst engaged in activities that will be compensated.”
Or did you expect developers to purchase the commercial version of Qt for this project?
If you need to have a toolkit then that’s halfway decent to program with and is LGPL’d, then use FOX and scrape together a combination of other libraries for network abstraction and whatever else. While FOX isn’t perfect, it definitely is useable and fairly robust.
I like this idea! Watson sounds like a very usefull app, and it would be nice having it on other OSes.
I’m not a programmer, but I might be willing to contribute a little to this project. it is important that applications that are easy to use, useful and worthwhile be a part of the open source movement.
About Java…again I am not a programmer, but Java is beginning to come around. I use both LimeWire and ThinkFree Office, which are both Java. In both cases, it takes them a second or two to open, but once open, they run fine. I especially like ThinkFree Office, it really has some nice features.
Supposedly, Sherlock 3 on OS X 10.2 is almost identical to Watson. It is unclear to me if Apple is working with the Watson people on this or not.
Why not offer $100 as the starting point for an original app instead of cloning someone else’s hard work?
How come Qt has better native support? AFAIK although it EMULATES the native widgets – it does not actualy use them like wxWindows does.
Hello,
It appears Watson is very smooth looking, but the Mozilla/Netscape sidebar applications do a lot of these things. How about writing tons of sidebar applications and hosting them on the Netscape website? You write these things in html/css/javascript/etc., whatever Mozilla supports (MathML and SVG in the future as well). It may not be “quit” as slick as Watson, but a completely different XUL app could take these sidebars and turn them into a very Watson-like application. Its also very portable (XUL/html/css require no recompiling across different platfoms). Just a thought!
–Wes
I guess there is something wrong with me – I see all this project other way.
Here is how Eugeina put it (in my translation):
“I want to be a project leader for this great idea. These are my rules. As I cannot contribute as a coder, I put $100.”
Flame me, as you want, but if it’s open source then the one who start actual coding will choose API, name, license and the other stuff.
And as Matt Kimball pointed out, receiving even $100 compensation would be violation of Qt free license.
Free ad space is actually another way of compensation too.
There is a subtle difference between this proposal and BeUnited driver fund – “rules”.
Drop the rules and then I click on that Watson URL.
Java has some nice features that would aid development – dynamic class loading *really* helps plugin development [been there, done that], great XML/XSL support, plus networking is a snap.
I think computers are fast enough these days to support the overhead of it running in a JVM, applications simply dont need a multi-GHz processor!
In a networked application like this, the network / file IO would by far swamp any performance hit from the language used. IOW, the IO would be the bottleneck, not relative speeds of C++ or Java. Each language has its merits but I would bat for the Java team on grounds of immediate portability.
As for license? BSD-style has a nice feel to it.
Speaking as a developer, I think this is definitely a Java kind of thing. It is best suited to the cross-platform development requirement, and has no “strings” attached like QT. Java could also be used as the back-end, so a developer could work front- and back-end without a language switch.
Any chance for a switch to Java?
I use TWM as my window manager because:
1. It is included with X.
2. No bloat.
3. I mainly use X just to get nice, large XTerms.
The GUI apps I use, besided XTerm, are xv, xbiff, and oclock. So anything developed would be useless to me unless it used the generic X toolkit or has the option of doing so.
> I want to be a project leader for this great idea.
Go and put yourself under cold water. I want NOTHING to do with the project. I just want a multi-platform app that can do such things. Therefore, I will pay my $100 as long as there are some rules to be met.
And no, I don’t want it to be Java or Kylix. Qt interfaces are cuter than wxWindows (plus wxWindows do not have OSX support), while GTK+ is simply not an option.
So, if someone wants to write the app in Qt, or want to write *native* versions of the app for each OS, there you get my money. Otherwise, I simply have no interest paying for a Java app. Even if it was a company releasing such a project in Java, and even if it was better than Watson, I would still not pay for it. I do not like desktop Java apps, and I would never pay for them, no matter who wrote them. It’s as simple as that. As for Kylix, it is not cross platform enough. I want with a simple recompile all the Unices to have access to the app.
http://sash.alphaworks.ibm.com/
http://oss.software.ibm.com/developerworks/opensource/sashxb/
SashXB for Linux is a “from the ground up” rewrite of Sash targeted at the Linux desktop. It uses Mozilla’s HTML layout engine, JavaScript interpreter, and Component Model. Sash also uses many components from the GNOME project. Two summers in development, SashXB is a product of IBM’s Extreme Blue internship.
I recently switched to Linux from Mac OS X. I loved Watson, but honsestly didn’t use it all that much, being shareware and all.
I have the skill to be a part of such a project. However, I’m not interested in having to put up with C++ and learn Qt to do it. While C++ is the tool of choice for many people, I don’t see anything inherent to this project that makes C++ a better choice. Frankly, the lower-level nature of C++ would hint at a higher-level language having the advantage. I mean, Watson is basically a GUI wrapped around HTML parsing. Perfect for Smalltalk, Perl or SNOBOL.
This makes me want to write my own version of a Watson world-a-like in Squeak Smalltalk or Perl (with GTK+, Qt, or Tk for a GUI). Luckily, both Squeak and Perl have no cost like Qt, and no license restrictions like Visual Studio. Perl runs most places, and Squeak runs on even more platforms- hell, you could even get native (no X11 required!) support for BeOS.
If anyone is interested in working on a Squeak version, email me. Drop the NO SPAM out of my email address.
But I know these kinds of opinions aren’t terribly welcome in a C++ world and aren’t very productive, so I’ll shut it lest I sound like a know-it-all language bigot and troll. All the same, I wish whomever takes this on the best of luck!
Eugenia, what you got against Java?
I just do not like the slowness. I have here an Office suite written in Java (ThinkFree), I got Jedit and a few other Java apps, and they give me the creeps each time I use them. They are extremely unresponsive and slow to load on my dual Celeron 533. It is like using OSX all over again, on Windows and Unix.
The other problem I have with Java, is that they do not look or even behave native enough, no matter Swing’s effort.
I would not pay $100 for something like it.
Other than these facts, I have nothing against Java.
Another vote for the name Moriarty here. Sorry Rajan R. ;-)Karelia seems to have chosen the name “Watson” because Dr. Watson completed the team of Sherlock Holmes and Dr. James Watson (as Watson is a “more complete” implementation of Sherlock’s search service), only no one knew that Watson would be superior to Sherlock in several ways.
Moriarty works because in the stories, Professor Moriarty was Sherlock Holmes’ deadliest and most brilliant opponent. The same could be said for any GOOD cross platform competitor to Karelia’s Watson and Apple’s Sherlock. It may at least motivate them to port it to Windows…
–JM
Even if I don’t know much about Watson, that project seems to be a good idea… Here’s some of my tough.. (PS: I’m not a developer, but I used to support 600 developers under different environment).
As for the developement environment, I you *personnaly* go with Java. When we talk of multi-platform, I think that Java may be the right one. If I understand clearly what i’ve read, you don’t have to care about the “OS call”, the JVM does that for you. Also, Java seems to be a good choice for network software. And, it seems that JDK 1.40 is much snappier on SWING.
Watherver language you choose, I approach like MVC (model – view – controller, often used in WEB Developement) might be good… So that people can re-use the core of the software and put the GUI (Qt, Gtk, Java, name it) they want.
Just my tought …
best regards and best of luck.
Martin
How about using GCJ to compile the Java bytecode to run natively on the target platform? This should provide a considerable performance improvement over interpreted Java. The big drawback is the GUI class libraries are not fully implemented yet…
`http://www.gnu.org/software/gcc/java/
LOL Eugenia, I know what you mean about Java applications giving you the creeps. I use ThinkFree and LimeWire on OS X. I launch them and wait…and then they sort of explode on the screen. You get the feeling that they’re about as stable as the San Andreas Fault and that they’re just hanging by a thread. But, I have to admit, they don’t crash on me. Somehow they hang in there. And on ThinkFree, connecting to my cyberdrive is almost instanteneous, especially compared to Apple’s iDisk. LimeWire also works very well. There is something about the way Java applications start up and, sometimes, act that is very different. It’s hard to pin down exactly what it is. I feel bad though, you made this proposal and everyone wants Java, the language you like least of all of them! But, I still think the project is a good one, an excellent idea.
What’s the difference between Sash and SashXB?
“Drop the NO SPAM out of my email address.”
Your a frick’n genious.
Swing doesn’t run well on slow computers (200<=) but it is a robust API. Dispite your feeling toward java, you got to admit that java web start is preaty damn cool.
Why don’t you all save yourself the hassle, pool up some cash, buy the right to port the software and give the poor developers their dues. If I were Karelia, I would have copyrighted the concept/technology/IP, and would sue pants off of anyone who tried to undermine my position. You will litterally be stealing the money from them.
Well, is Karelia had Linux and Windows versions, there would practically be no such proposal, right?
If this goes ahead, and uses anything less than the GPL, people could proffit greatly from it, and that’s just plain wrong. It’s a sad world when the only good ideas are those we steal from legitimate business for our own gain.
Without allowing businesses to use it is like removing the functionality of the application itself. In case you didn’t notice, Watson depends a lot on their servers, and possibly pays a lot to their partners so they could get content to Watson. Without the content, Watson is useless. If nobody could make money, for sure nobody would throw away money for stuff like bandwidth cost, and content partnership to get a server to provide for the users.
Why not offer $100 as the starting point for an original app instead of cloning someone else’s hard work?
Get a life, troll. The end product may not even be like the original Watson. Eug is just copying the ideas of its function, but for stuff like content presentation, the type of content, the user interface and so on may not be the same. If everyone was to follow your philosphies, there would may never be most of the worlds OS and applications; including leading ones. Remember, Excel was a Lotus clone; Word was a WordPerfect clone.
It appears Watson is very smooth looking, but the Mozilla/Netscape sidebar applications do a lot of these things. How about writing tons of sidebar applications and hosting them on the Netscape website?…….
The sidebar idea is kinda good; but it not the same idea as the Watson one.
Flame me, as you want, but if it’s open source then the one who start actual coding will choose API, name, license and the other stuff.
Since she started the idea of an open source clone (or one that is daring enough to say so), and also she is going to invest a lot into the project. So, I think she deserves the right to to decide the product outcome.
And as Matt Kimball pointed out, receiving even $100 compensation would be violation of Qt free license.
Free ad space is actually another way of compensation too.
Which makes KDE e.V., and KDE League illegal too?
The GUI apps I use, besided XTerm, are xv, xbiff, and oclock. So anything developed would be useless to me unless it used the generic X toolkit or has the option of doing so.
Seeing the list of applications you claim to use; this is no use to you, or at least no made for you.
I recently switched to Linux from Mac OS X. I loved Watson, but honsestly didn’t use it all that much, being shareware and all.
Did you stop using it because you don’t need it or you have to pay $30(?) for it?
Another vote for the name Moriarty here. Sorry Rajan R. 😉
Damn! I demand a re-election!
I like Moriarty also…
Swing doesn’t run well on slow computers (200<=) but it is a robust API. Dispite your feeling toward java, you got to admit that java web start is preaty damn cool.
I don’t see how Swing is more robust that any of the suggestions thrown in here (e.g. Qt, wxWindows, FOX…)
I look at the feature list of the suite; it is not worth it for $50. Just add $25, you would get something at better features and more compatiblity with Office. Plus, you can get all the features they rave in ThinkFree (except the Java part) plus more. Plus, ThinkFree isn’t that portable after all, as there is only Windows and Mac OS releases!
I was just googling to see if I spelt Moriarty right,
Amazingly there is even a fine fellow Dr D Moriarty down in TX works for Apple with that very name, & is even possibly qualified to write such a program, as resume lists “data mining, Java,..C++,… AI, etc etc. I wonder if he was called Professor in jest.
I agree with Eugenia in regards to Java. I would also like to remind people that not every operating system has Java.
> I want with a simple recompile all the Unices to have
> access to the app.
Not realistic. Being a developer – one who has actually done some cross platform work and also ported code from platform to platform, I can honestly say that the above is *not* an option. If you mean ‘one makefile’ then, yes it’s possible. But then that’s possible with Kylix/Delphi/FPC too.
The best approach would be to write the cross platform sections engine completely in vanilla C. The GUI could then be written in anything, and there’d be a much higher chance of cross platform recompilation. It would also open up the project to PDA ports. No way will a C++ monster run on, say a Palm. That would be where it would be really useful.
Getting paid to work on this app would mean that the developers would have to pay the Qt License. It’s not a commercial app. If the app is open source and not sold on a binary-only basis, it counts as non-commercial. So quit being assholes about it.
I’m intrigued.
I’ve written something vaguely like this, although it’s for browsing rss links (I actually use it to get the latest OSNews info and display it in my browser. Thanks for the RSS feeds, BTW) and to view my favorite on-line comics. I’ve got a plug-in architecture on the drawing board, as well as some other good things you don’t usually see in a user interface.
Let me ask you a question:
My application is written in Python. I use PyGtk (Python + Gtk) on my machine, though I’ve also ported to Win32(using Tcl/Tk. I know, it’s yucky, but my friend didn’t run linux, so what was I to do?). Simply replacing the Gtk calls with Qt calls in the linux version would be trivial. For a nicer looking windows version, I could use PyQt for windows (which I believe is still a little flakey) or WxWindows (which should work pretty well). I think WxWindow bindings for python and MacOSX are in some kind of alpha version.
While I don’t have anything against C++, python would make portability a hell of a lot easier and development a heck of a lot faster. If I could deliver said app in python, would that be acceptable?
> What we really need though, is willing and capable C++ developers.
IMHO this is incorrect. software (an the development thereof) would be in a much better position if c++ (and the likes) could be replaced by one of the simpler, more high level, more secure, etc alternatives.
bengt
If this ever gets past people arguing over the language, java comments, and the qt licenses some progress might be made.
I think it sounds neat, any private respondants taking up the challenge?
Again, I’m interested in working on a project like this in Squeak. I’m serious- if there is anyone out there with any Smalltalk knowledge and would like to work on this let me know, I can see a very useful tool coming out of it. Plus, it’d be incredibly cross-platform, running on damn near anything.
The whole idea of the project is to be largely multi-platform (note: Java is not an option – read my previous comment).
If your favorite language/toolkit is not as such, then it is not an option, as far as I am concerned.
I know that you’re set on C++. People big into C++ are just as hard to bring to reason as anyone else. However, I’m not talking about getting you to change the “rules” of this project- I’m just putting it forward that I’m still interested in doing some real work on something like it, done in Perl or Smalltalk. Rather than just bitching about how much C++ blows.
RevAaron, Ilan Volow,
Do your own thing, you are the boss of your own time & dev efforts unless someobody gives you a paycheck!
The only concession a developer must make is if they need to find other like minded developers with same common tool interests.
Go for it!
Eugenia wants a C++ app. If you want to use Smalltalk, you could use Smalltalk, but not the $100 and the ads. Maybe, just maybe, Eugenia falls in love with your apps and bend the rules, and then viola you get the benefits. Otherwise, this is Eugenia’s money and part of Eugenia’s site.
Heh. I know. I’ve said a few times that this would be seperate. I don’t expect money, especially not Eugenia’s. As I’ve said, I’m not arguing that she should change her rules. Just putting the idea on the table for anyone else interested.
IMHO this project would be a real “killer” if it provided
a (data) backend for use by other applications.
There are lots of programs that have interfaces to web applications, just some examples that come to my mind atm:
– Plugins for Abiword (like Babelfish)
– the Evolution Summary
– Emacs (Wiki, I’m sure there is also a Babelfish script)
– GNOME/KDE panel applets
– WindowMaker applets
– cgi Skripts that use data from different sources in the web
– all kinds of web agents
– Bookmarklets/”smart bookmarks” for Galeon, Netscape, IE
Whenever the web service’s interfaces change, all the developers have the same problem: They must change the program or their program data.
I suggest to design an XML DTD that is able to describe the necessary interactions between the web service and a client program.
For every service, there will be an xml file.
It does not need to be limited to http, the language could be enhanced to support E-Mail (esp. mailing lists, one mailing list manager <-> one file), the dict protocol (I don’t know if this is still important), LDAP, public key servers, and so on.
Application delvelopers can then use this data files in their programs. If they want to add new services, they can do it for their applications (and hopefully contribute it to
some sort of central repository).
If someone does not like XML, (s)he can translate the files
automatically to another format …
The applications themselves could be written using any language and any toolkit (C++ and Qt in your case, Eugenia).
regards,
Zeno