Latest News from the Crumpery

Tasty

Mon, 15 Dec 2008 22:34:19 +0000

Tasty is my first commercial application for the iPhone and is currently being prepared to go on Apple’s App Store.

So what is Tasty all about?

Tasty allows you to rate and keep track of your favourite (and not so favourite) restaurants and other eating establishments.

You can rate the meals you’ve had, the ambiance of the establishment, quality of the service and overall cost of the meals.

Tasty will be useful for your holidays and business trips where you can tell your friends or remind yourself which places were good and which to avoid.

You can associate pictures you take on the iPhone with the restaurants themselves, the meal and even the individual courses.

Version 1.0 concentrates on 'the gathering' of your data. Future versions will improve searching and sharing of the information you capture.

Watch the App Store for it’s release.

Distillation Designer Breaks 2000 Downloads

Sat, 11 Oct 2008 14:25:03 +0100

Distillation Designer continues to be downloaded and has exceeded 2000 downloads (based on VersionTracker, MacUpdate and Softpedia counts.

Still no comments have been received so I would love to get some feedback on anyone who finds the program useful.

In the meantime, my iPhone application is nearing completion and will soon be released into beta testing.

I’ve also got myself an iPhone contract writing an application to read and display real-time data feeds from a server, absolutely fascinating stuff and loving the work.

My iPhone Program

Sun, 7 Sep 2008 23:39:45 +0100

My iPhone program is coming along nicely (a few more weeks work and the bulk of the work will be done), I am currently waiting on Apple to register me so the process of getting the program on the App Store can start.

All I can say at the moment is the program will allow you to store and review items you come across. The concept is not ‘earth shattering’ but I hope it will be useful addition to the iPhone/iPod Touch for people who travel a lot.

It has given me a good understanding in developing forms using the Cocoa Touch API. So if this does not sell, I might still get an interesting job developing for the Mac and iPhone in the future.

Fingers Crossed...

RETROCHALLENGE 08 - Final Day - Reflections

Thu, 31 Jul 2008 09:01:03 +0100

This Retrochallenge has been very useful for me and I’ve got two of my important machines in my collection back to life and/or integrated into not just my network, but the internet as well. Both machines are safely backed up, configured and have some very interesting (and potentially useful) software installed ready for further investigation. I’ve done everything I set out to do.

Before the challenge, I was looking forward to getting back into Open Look/Solaris and perhaps do something like a little XView program. Instead, my journey took me to NeXTstep not only on the NeXTstation but also on Solaris (through OpenStep). This was something totally unexpected, I never imagined I would be able to see Solaris and NeXTstep/OpenStep in colour first hand. It has been truly the highlight of the challenge for me.

In fact, I am starting to think that dumping OpenStep on Solaris before it really got chance to flourish was probably one of the worst decisions Sun Microsystems made. After all, what did moving to Motif/CDE really get them apart from a standard most UNIX vendors wanted to get away from once Windows 95 scrapped the Common User Access interface style! Instead, Sun would have had a desktop environment which was modern, clean and powerful with a fantastic API in which to develop with.

Why did Sun reject OpenStep at such a late stage? Was there a power struggle between Motif and OpenStep camps inside Sun? How much did Java (promising to do everything at the time) affect the decision? Did they not trust NeXT or did they believe NeXT was heading for bankruptcy? If so, why didn’t Sun just purchase NeXT before Apple did? Did Sun really believe a unified UNIX desktop using Motif (with it’s (by then) out of date interface style) was more important to pushing UNIX desktop market forward? Lots of questions and no answers at the moment, perhaps you know more? At the time I remember the onset of Java was used to stop the development of a number of very promising technologies including OpenDoc at Apple, where they ‘explained’ how Java would replace the need for OpenDoc. I never bought that idea then and well, it never did!

During the time of Sun dropping OpenStep in the mid-to-late 90’s, I remember not being concerned about its implications since I never had the opportunity to try it out on a Solaris system or in fact any system. NeXTstep always remained an illusive technology to me and I would think many other UNIX developers at the time. Sun has now moved into the open software camp with their Gnome desktop (copying, in a second rate fashion I’m sorry to say, the look and feel of Windows 95/XP). Their design simply does not impress me like NeXTstep does even though we are now 16 years on since its (v3.0) introduction.

I wonder how things would be now if Sun did follow the NeXTstep/OpenStep route. Would we now have SnApple (Sun/Apple), both organisations bringing their strengths, with Sun on the server side and Apple on the desktop and joined by a common API and desktop? Even without merging, perhaps Sun would still be strong in the desktop game utilising Apple’s desktop strengths, while Apple having a much stronger business presence through Sun’s influence in the server marketplace.

Now those are some thoughts I wouldn’t have without doing the Retrochallenge! :-)

I’ll look forward to next years challenge where I just might go further retro and write a SuperBasic program on the Sinclair QL. But what to write? hmmm. Well I’ve got a year to think about that.

RETROCHALLENGE 08 - Day 30 - Trying out NeXTstep Software

Wed, 30 Jul 2008 19:40:12 +0100

Compared to Solaris 2.5, there’s still an awful lot of very good quality software available for the NeXTstation that comes with an ‘open’ license. I was actually surprised to see so much, especially since the NeXT computers sold a lot less than Solaris desktops at the time.

Here are a few highlights:

FrameMaker 3.0 (see above).
This is just a demo version, still, I have used FrameMaker on both Solaris and Macintosh and this version looks great, especially in a Display Postscript environment. I think I’d be very happy to use this today, if I could get a license. (I wonder if my Macintosh license would work in this, probably not... Ho Hum...). 
Diagram.app 
With this high resolution monitor, this looks very nice application, again it would not look out of place today!

plot3d

This does render slowly these days, but considering we are on a 68040 CPU, I’m impressed by it’s speed compared to other machines I used at the time. 
EqB 
Now this might be useful one day, being able to create equations and exporting a Postscript version to my Mac ;-)

OmniWeb

Of course this application struggles with modern javascript based web pages, but at least I could not get it to crash, just weird rendering at times. Still another success for the Retrochallenge web page. 

Lotus Improv

I bought this copy a couple of years ago. I just love how neat this spreadsheet looks compared to modern day Excel, it reminds me of Apple’s iWorks Numbers, though probably Improv is still a lot more powerful than Numbers.

Missile Command

This was fun, though the thing which impressed me the most was the sound quality. There appears to be a very good sound processor inside perhaps it’s the DSP which is being used? 
There is plenty more software which is free for NeXTstep which I will need to try out, sadly time is against me at the moment. So far, all the software I played with was such good quality, it makes me wish I spent my early years using a NeXTstation instead of Solaris! One thing which does grab me is the simplicity of all the interfaces these NeXTstep applications had, yet their power was always close to hand. I really like the design concept of these applications. They are very cleanly built by people who appear to really understand good user interface design.

I think I need to spend a lot more time with NeXTstep, perhaps this will be a RetroChallenge for the future.... In the meantime, I think I’ll keep a look out for a colour NeXTstation, mind you, the last one on ebay (UK) went for over £400 ($800), it appears there are other people out there who know how impressive these machines really are!

RETROCHALLENGE 08 - The Case for Open Look

Wed, 23 Jul 2008 13:24:57 +0100

On Day 2 of the Retrochallenge I said I should explain why Open Look was so good so here it is:

In the late 80’s, there was attempt by a number of companies supporting UNIX to standardise on a single code source for UNIX with a single GUI (Graphical User Interface).

Sun teamed up with AT&T to create Open Look which was based on the X Windowing System as well as a new version of UNIX (System V Release 4). Unfortunately other companies, afraid of the power of Sun and AT&T at the time, ganged up to create a rival competing UNIX and GUI called OSF/1 and Motif respectively.

In Open Look, different shapes within the GUI provide a powerful set of visual clues for different interface functions. For example:

Buttons have rounded corners.
Popup Menus are associated with indented triangles.
Pull down menus are buttons with an indented triangle.
Sliders are three squares bolted together (two with triangle shapes) attached to a ‘sliding rail’.
Toggle boxes are square with a ‘tick’.
Radio buttons are horizontal rectangles.
Window controls are square with an indented triangle inside.
Drop sites are rectangles with a patterned ‘document’ inside.

While In Motif:

Buttons are rectangles.
Popup Menus (combo boxes) are rectangles with a smaller rectangle inside.
Pull down menus are rectangles.
Sliders are rectangles inside a bigger rectangle.
Toggle boxes are square rotated 45 degrees.
Radio boxes are horizontal rectangles.
Window controls are square on one side and square with a rectangle inside on the other.
Well, you get the quadrilateral gist...

The main design constraint of Motif was to follow the Common User Access style guide (which Microsoft Windows 3.1 and Presentation Manager interface from OS/2 also followed), this is probably why Motif lacked imagination in it’s design!

The use of a ‘drop site’ in Open Look also helped make the GUI much more ‘object based’ than Motif, since you could drag data (in the form of objects) directly from one application to another. (Below is an example of this where a ‘component’ in one program could be added to the ‘stream’ (containing many components) in another using the drop sites at the top right hand corner.

Open Look had two API’s in which to develop with. OLIT (from AT&T) and XView (from Sun). OLIT was similar in design to the Motif API being based on the X Intrinsics API. XView originated from Sun’s earlier SunView API.  
I found the X Intrinsics API while highly configurable, overly complex. In 1989, I remember getting the company I worked for to purchase the DecWindows programming manuals. DecWindows API was the forerunner of the Motif API, their manuals cost £600 and after reading them I still found it impossible to write a Menu in a Window such was the complexity for doing something which should be so simple. Documentation for Motif did get better, have a look at Programming in Motif which is now available for free on the internet. However compare how you create a simple popup menu in Motif with similar code in XView (see page 280-281). Xview has five main functions (xv_init, xv_create, xv_set, window_fit, window_main_loop) to create and define the window and menu while Motif needs 9 many of which are actually direct X Intrinsic calls (these are the ones beginning with Xt) (XtSetLanguageProc, XtVaCreateManagedWidget, XtVaAppInitialize, XmStringCreateLocalized, XmVaCreateSimplePopupMenu, XmStringFree, XtAddCallback, XtRealizeWidget, XtAppMainLoop).

Open Look also came with a free and easy to use GUI builder (DevGuide), while most companies developing in Motif used a very sophisticated and expensive GUI tool like UIM/X.

For more of an in-depth background on the design considerations of Open Look, there’s an excellent article here.

Case Closed.
Alas, although Open Look had a more intuitive graphical design and an easier development environment they were not enough to stop (commercial) developers building around Motif since more companies and therefore more versions of UNIX were supporting that standard. i.e. Why develop for Open Look on Sun when you could develop in Motif and provide a version of your software to many more companies UNIXs (Unices?) in a single ‘hit’? Eventually Sun had to cave in to using Motif/CDE as well with Motif going standard on their desktops around 1995/6.

So where did a unified UNIX desktop get us? These days the main OSF/Motif partners (in terms of volume) IBM, DEC, HP and (one can argue) SCO, barely mention Motif desktops in their websites. I understand two of the original partners, HP and IBM, did not even move to the OSF/1 version of UNIX! DEC is now part of HP and you have to dig deep in the HP website to find (desktop) HP-UX (the origin of Motif’s 3-D look and feel) mentioned in any great detail, while IBM is more interested in selling you Linux desktops than their own desktop version of AIX.

Although OSF members may have won the UNIX GUI battle, one reason why I think they lost the desktop GUI war is that Motif was simply not a very appealing GUI to many prospective users.

For a while, Sun toed the ‘party-line’, but I felt somewhat grudgingly since they finally stopped including Open Look (Open Windows) on Solaris in 2002, they have also since abandoned CDE/Motif for a new desktop based on the Linux GNOME API called the Java Desktop. The Java Desktop resembles other Linux derived desktops which in turn continue to have a strong Windows flavour.

Copycat GUIs.
I find it interesting how (open software) interface developers believe if they make their GUIs resemble Windows, users will be more inclined to use it. Instead, the most successful UNIX based desktop GUI in terms of volume is not Motif nor Linux based but something with its own graphical style. i.e. Mac OS X. Instead of users thinking a copycat GUI is familiar, I believe many are reminded of Windows thinking the Linux GUIs are just inferior copies. After all, when was the last time you thought a copied product was better than the real thing?

To get around this, you can see many applications (supporting these copycat interface styles) are ‘prettified’ with icons which actually just clutter up the GUI (especially when added to menu items and buttons). Perhaps instead of trying to create a copycat GUI, someone should review the Open Look desktop to get an idea that their designs could be cleaner, more intuitive and perhaps more appealing if they actually dared to be different and more importantly, really innovative.