ROX founder: Why I brought RISC OS to UnixBy Nick Brown. Published: 25th Jul 2007, 00:56:04 | Permalink | Printable
Making Linux easy to use, advice for ROL and CTL, and moreDrobe's Nick Brown has been granted an exclusive interview with Thomas Leonard, the creator and visionary behind the ROX desktop. Here, Dr Leonard answers questions on ROX, how RISC OS inspired him, what he thinks about the current state of the OS, and what the future holds for his desktop.
According to rumour, the start menu and task bar in Windows was included after Bill Gates used an A5000 - who knows if that is true, but the design of RISC OS has inspired many developers way beyond the platform itself. One such developer is Thomas Leonard and the open source ROX (RISC OS On X Windows) Desktop project.
The aim of ROX has always been to combine the best features of RISC OS with those of traditional Unix desktops, including Linux. Its ever growing user base has meant that many people who have never heard of RISC OS can experience aspects of the operating system's design which have proven to be popular. Essentially, ROX acts as a desktop environment within the X Windows system, often found running on Unix-like computers.
A typical ROX desktop. Click for larger.
Drobe: Firstly, please tell us a little about your RISC OS background - how long have you been using RISC OS and what was it that that attracted you to it?
Dr Leonard: "I started with RISC OS 2 in 1990 on an A3000, upgrading from a BBC B. It was a pretty easy choice back then."
In total how many active developers are there working on ROX - how many of them would you think have used RISC OS?
"It's hard to say. There are 160 people currently subscribed to the developers' mailing list, with perhaps 20 or so people actively sending patches and writing code over a long period."
Broadly, what was it about RISC OS that inspired you to write a new desktop? What features in particular did you find lacking on other desktops?
|"I'm impressed that RISC OS still exists. It does sometimes get mentioned on news sites, and the question that is always asked (and never answered) is 'why would I want this?' What is the vision for the future? Castle and ROL need to communicate the advantages more clearly."|
"The main ones were: a filer-centric interface; drag-and-drop everywhere, especially saving; application directories and a simple file-system layout; and decent font rendering."
How much of the current ROX is based on RISC OS? As you have developed ROX, and features have been enhanced, do you feel it has grown away from the initial brief of 'RISC OS On X' at all or do you keep a tight control on the development? Apart from RISC OS, where in particular have you drawn design inspiration from?
"ROX has always been about combining the best of Linux and RISC OS, rather than simply recreating RISC OS. The main features from Linux I've tried to include are: a general focus on security everywhere, for example, no !Boot files; the powerful command-line interface and scripting support; and keyboard-shortcuts through-out.
"We've tried to take concepts rather than details. For example, we use pop-up menus, as in RISC OS, but you open them using the right button by default, not the middle one. We write most of our programs in an interpreted language, but we use Python rather than BASIC."
From a purely interface point of view, do you feel that a RISC OS user could easily migrate over to ROX, and if so, what do you think would they benefit from the most from?
"Many already have. The first thing you're going to notice is the huge range of software available at no cost. Applications may not all follow the ROX style, but they often integrate pretty well. You can have any non-ROX application on the iconbar and drag files onto it to load them, for example."
The ROX desktop will appear familiar to RISC OS users. Click for larger.
A major feature of both RISC OS and ROX is that applications are directories. This affords flexibility that many desktops do not. As people consider this both very useful and natural - do you feel that other desktop developers have missed out on the idea?
"For a long time there were arguments about which way was better. The RISC OS way puts the user in control: when you install an application, you know exactly what happened to your system - that is, a new directory was created where you dropped it.
ROX application directories. Click for larger.
"The traditional Linux way is to use an installer that downloads the program and any required libraries and then scatters the files all over your disc. It's quicker and easier, and makes upgrading simpler, but you lose control. All kinds of strange things can happen. For example, I installed one program recently that, as a side-effect, changed the background image of my login screen.
"For ROX, I created a hybrid system called Zero Install. It has the same properties as plain application directories, which ROX also supports, but with automatic dependency handling, security checks and updates. That is, you can ask to install Edit (the ROX text editor), and it will also install ROX-Lib, which is a library Edit requires, like the SharedCLibrary or the Toolbox on RISC OS. But it will do so in a controlled and predictable way.
"I'm trying to build this to be useful beyond ROX, so that when other people use it to distribute their programs it becomes trivial to convert them into ROX application directories automatically. You say to your computer, 'Get me the application on that web page as an application directory,' and a save-box appears letting you save it, even if the author of the program has never heard of application directories or ROX.
"For example, you can drag Xara Xtreme (the new name for Artworks) from this web page to ROX and get an application directory ready to use, even though the authors of that program never designed it for ROX."
Are there any features of RISC OS that you would still like to implement in ROX?
"There are still lots of Linux applications that don't do drag-and-drop saving. On the other hand, the font situation improved massively a few years ago so that's no longer an issue. That wasn't due to our efforts, but that's the point of ROX: to take advantage of advances in hardware and software going on in the mainstream computing world without extra effort on our part."
Drag and drop saving on the Unix platform.
One of the comments that people mention about ROX is that, even though the desktop and its applications follow the design principals set by yourself, all the other system software follows a different design. Is there any way that this might change at all, or will applications need to be rewritten in order to follow the ROX design and employ features such as drag and drop?
"The main ROX feature missing from most applications is drag-and-drop saving. Actually modifying any individual program to support it is easy, perhaps the work of an afternoon, but then you have to keep releasing new patched versions as new versions are released by the original authors, or the ROX version quickly becomes out-of-date.
"Ideally, applications shouldn't need changing at all. An author should write a program and publish it, and a user should say how they want their programs to behave - for example: 'I want application directories and drag-and-drop saving.'
"Without a large body of users asking for that kind of flexibility the people writing the core libraries probably aren't going to do anything, though."
Have you kept up with the RISC OS scene at all - what do you make of its current state?
"I'm impressed that it still exists. I do wonder what might be done with ROX with similar levels of funding, though... one full-timer can make a lot of difference."
If you could give RISC OS Ltd or Castle some desktop design advice, what would it be and what feature do you feel is most missing from RISC OS?
"RISC OS does sometimes get mentioned on news sites, and the question that is always asked (and never answered) is 'why would I want this?' What is the vision for the future?
"Riscos.com says they have a 'goal of releasing a full 32 bit version of RISC OS within 12 months'. It's hard to see why any new user would care about that. Castle's site lists a few things, such as 'anti-aliasing, Unicode font handling,' but nothing that other systems don't offer too.
"They need to communicate the advantages more clearly."
Lastly have you any major new features planned for the Desktop?
"Right now, I'm mainly focussed on the installation system and getting more applications using that.
"Thanks to its integration with Linux and Unix, though, almost every new feature planned for Linux also applies to ROX. For example, there are people working on using 3D acceleration for transparent windows and special effects, others working on new security technologies and virtual machines, longer battery life for laptops, support for even more devices, new types of file-system, improved programming languages, better web-browsers... the list of exciting new stuff is endless."
ROX website - what is ROX?
Zero Install website
All images courtesy of the ROX website.
Previous: Free virtual memory app for CD mag readers
Next: Peter Bondar speaks: He heads software company
DiscussionViewing threaded comments | View comments unthreaded, listed by date | Skip to the end
Please login before posting a comment. Use the form on the right to do so or create a free account.
Search the archives
Today's featured article
Oregano future undecided
Meanwhile, TV Browser reaches version 3
39 comments, latest by nex on 09/12/04 2:46PM. Published: 25 Nov 2004
Could open source RISC OS bring back users?
The jury deliberates
27 comments, latest by scl4c0rn on 15/12/06 9:48PM. Published: 10 Dec 2006
News and media:
RISCOS Ltd •
RISC OS Open •
MW Software •
Advantage Six •
CJE Micros •
Liquid Silicon •
Chris Why's Acorn/RISC OS collection •
The Register •
The Inquirer •
Apple Insider •
BBC News •
Sky News •
Google News •