Continue reading »]]> I’m still working on it… still in the process of implementing file management features, mostly graphical elements that allow for handling of files. Like list them in list view or icon view, etc.

It is gonna take a bit more time on that front as I thought it’d take… experimenting around with NSTableView – which I find very very difficult to handle. Anyways… The thing is still progressing and has not come to a halt.

Continue reading »]]> In the last few weeks I have been working pretty hard on my project, which eventually is meant to become a FreeBSD based operating system, with all likelihood commercial in nature.

Recently the entire code base has crossed the 5,000 line mark, and it’s all working out nicely.

I can tell you that I have developed features that currently are not to be found in all modern operating systems, like Windows, Mac OS X or desktops on Linux (KDE, GNOME, etc).

I cannot tell you the specifics about it or post screenshots because I want to keep these features a little closer to my vest in the moment.

Rest assured that it is progressing nicely and I am now in the area of implementing file management features.

For now, you can have a look at the placeholder page I have put up at the domain for the project, laststep.eu.

For the technical people out there:
laststep is built upon a FreeBSD 7 core, which then has the GNUstep libraries and frameworks – upon which my code is built on. Some of the features use GNUstep quite extensively, while others built upon my own frameworks (which then in turn use GNUstep).

GNUstep is a cross-platform, object-oriented framework for desktop application development. Based on the Cocoa API (Formerly OpenStep) originally created by NeXT (now Apple), GNUstep enables developers to rapidly build sophisticated software by employing a large library of reusable software components. GNUstep is used in production environments at several organizations.

Continue reading »]]> So I quickly want to keep you posted on progress on my project. It’s been a while since I worked on it, however I have been working on it the last two weeks. You see when I did the whole thing in Squeak, I made progress quickly – even had a working file browser. But one thing brought it to a halt – it is single threaded – that means that you could only have one application running in full effect where as the others would not be executed correctly. So Squeak… it was nice. But no. If you want to do your own thing, you have to run on the chip.

Because of that I have been starting to port my old Smalltalk code into Objective-C. First results below. Comparison between old and new code.

Click to enlarge. Current status with menu bar in Obj-C and new Drawer Concept.

Click to enlarge. Where COSMOS was when I left it.

As you can see, the menu bars are nearly identical.

Continue reading »]]> Alright. So I have had a good look into Objective-C for the last few weeks. I think I have learned enough to start converting my original code from COSMOS (Smalltalk) into laststep (Objective-C). It is not much, but I have been able to build my own desktop with wallpaper, and a first port of the menu bar. It doesn’t do much yet, but it’s a start.

The lastDesktop with a first test of the menu

The original design of the menu I have written in Smalltalk

I’ll update you on the progress.

Continue reading »]]> As some of you, I am working on my own kind of distribution. I have continuously posted screenshots on my page. For now, my main page is offline as I am restructuring information on it, I am also in the process of setting up an all new domain, devoted to the distribution project alone.

Some of you may also have gathered that not only I am working on my own distribution, but on my own desktop, or end user interface. It’s a big task all right, but it’s also fun to learn. In the end I’d be able to say that I learned these languages, but with this one I made something really big.

As I have progressed in Squeak and Smalltalk, and as I got down to serious business, I have noticed that there is a big slowdown happening, whenever you have more and more windows open, regardless if they have controls and elements inside them or not. Goes for Squeak’s SystemWindow implementation, as well as for my own framework inside which you can find CSMSWindow. It’s fine when you have one or two, but when you have like 5 or so, a slowdown is really noticable.

The Squeak VM running on X11

This has to do with the fact that processes of Morphic, Squeak’s current major graphical implementation, is not asynchronous. This means that the current active Morph (and thus a window) can have the processor’s full attention, while others stay behind. And this, my friends, is really lethal if you want to build your own end-user desktop interface.

So what are the options?

Contender #1: wxWidgets

Platform independet framework. Lets you write apps on almost any platform and then take it to another system with no or minor modifications. Many applications are written in it, mostly Linux X11 programs. The big advantage is its easyness, or so it is claimed by many. The disadvantage of it would be bearing no similarity to the code I already wrote. But then again if it is as easy as they claim, maybe it could be rewritten. Let’s quickly look at some code.

Creating a window takes some lines and 4 files. And the window has nothing in it.

While this is still impressive at the thinness of code, I was looking for other alternatives.

Contender #2: QT by Trolltech

Also platform independent, and has found roots in many things out there, most notably KDE. However, the site claims that Skype, Google Earth, Adobe Photoshop Elements and Lucasfilm are amongst their customers and use their framework in their applications. Well that surely sounds like something worth looking into, right? Let’s see how they do a little Hello World application, utilizing the framework. Turns out this is all you need.

Not so bad. Better than wxWidgets all right. 13 lines of code. Then run the QT compiler and it’ll do the rest for you. Awesome. Advantage appears to be that the code is amazingly thin, and can be used on all platforms. But again, my code wouldn’t be compatible, I’d have to start from scratch again.

There has to be something out there.

So that forced me to look, again. Look at possible alternatives, but this time I kept in mind that I know Smalltalk-80 pretty well by now. I have written my own UI framework in it. So. Smalltalk. It is object-oriented. And came from the early days. Is written in itself. All these things I kept in mind. After a thorough amount of search I came across something, which is the winner in this contest.

Contender #3: GNUstep

Platform indepedent. But who is using it, or what advantages does it have? I’m sure many of you out there have not heard about it.

It is important to understand where it comes from. Before Apple got to its glory it has today, there was NeXT, Inc. Founded by Your Steveness himself. They then developed an OS, named NEXTSTEP. This OS in turn had something no one else had back at the time – an object-oriented user interface backend and API. This means that back in 1992 when it came out first, it was so advanced, that you could compare it to Windows 2000.

It is also important to understand what made all that possible: Objective-C.

NeXT developed extensions to the C language, which would allow for the very same structuring and message parsing as it was already seen in Smalltalk. Objective-C in turn then used to build the NEXTSTEP user interface.

Shortly before Apple began developing Mac OS X, they purchased NeXT Inc. The base of NEXTSTEP then became the base of Mac OS X.

However NeXT made its protocols available to anyone, known as the OPENSTEP specification. And from that foundation, the GNUstep initiative was built.

What does this mean? This means: GNUstep shares the very same foundation as Mac OS X, while being platform independent. The language also is Objective-C, which is highly compatible with Smalltalk, as it served as the original for the new superset of C. Having that said, you could design and develop an application on GNUstep, then take the files to Mac OS X and compile it there – it will work.

This is why I am choosing GNUstep as my environment for expanding my code, and porting my code to it. It would also run directly on the chip of the machine, without a VM. But what impresses me the most, is that I can use the foundation of one of the most advanced systems in the world, and build my environment on it. In fact, on the development machine I set up, there already is a very similar environment to Mac OS X running, however for the moment I use WindowMaker with the NEXTSTEP look.

Let’s see what develops.