Just curious, thats if the Xojo team want to say, what percentage of computers at Xojo are Windows/Mac, and what do people prefer using?
the challenge is that people like myself that has a Xojo IDE on Mac, I write code for Mac, Windows, Linux, WE and RPi (so just about all of them). It would be hard for them to see what I am writing for unless they do something odd like keeping track at compile time of what I am compiling to.
I have a Macbook Pro and prefer working on it as a Mac, but I also have it configured for Windows 10 under Bootcamp and work try to work there at least a few days a month.
iMac 27" + MacBookAir, Parallels with Windows 7 and 10 on the iMac mainly to check the Windows builds are fine and cross-platform file-format compatibility.
Almost all of my programming is on Windows 10 and most of the programs run on a Windows machine at work.
Occasionally I get a request to run a program on a Mac, and just press a button to compile the program to run on OS X. I am noticing that there is an increasing demand for electronics programs on the Raspberry Pi.
I have both an iMac and a fast desktop PC.
I found a while ago that for complex programs, cross compilation becomes extremely complicated as the program grows, so now I have a version for each.
This is of particular importance on PC, where optimization to reduce flicker is not a joke. Working on the actual machine with the Windows IDE and the PC keyword let me fine tune the program so it perfectly conforms to the UI Windows users are accustomed to. This very morning I realized I had kept Option-Cmd-? as shortcut for help, which is quite fine on Mac, but irrelevant on PC where the expected key is F1.
I don’t think any serious developer should work on anything but the target machine, at least for the last mile.
For my plugins it is about 2/3 Mac IDE and 1/3 Windows IDE as far as I know.
Linux is hardly used for the IDE.
It seems that this conversation is going somewhat off topic and while @dave duke appears to be happy with the responses so far I’d be interested in the answer to the original question.
Wayne, i guess you know the answer: 100% Mac, for Windows only VM’s. I believe Norman has somewhere under a pile of dust an old pc. I think even those VM’s are all covered with dust.
Andre: I would not put money on the Table about that.
Well… Just a historical example… A large government customer for whom I wrote software way back in 2000-2001 - the early days of RealBasic - was at that time a Windows-only shop - and still is. They wanted to integrate real-time GPS data from approx 2000 trains state wide, sensors that weighed and identified each vehicle, a billing system that delivered invoices in near-real time using SOAP, and deliver a graphical web-based user interface showing the path of each train that serves circa 10,000 internal and external users. Unfortunately the developers originally contracted decided to bail out midway, after a million dollars had been spent on them at hourly rates. What’s worse, the lack of a working back-end implied contractual delays - and serious penalties - from other contractors delivering more systems that would interface to this.
To get the customer out of a very ugly commercial pickle, two of us solved the whole problem rather elegantly by wroting several apps using REALBasic on an iMac and PowerMac G4 running System 9. At that time the apps were written in RealBasic on an original iMac, tested and debugged with the iMac “in the loop” of the real system - ie with live data, then compiled for Windows and installed as services running on a PC running Windows NT in a server room. We also had a very good PC emulator on the PowerMac from Connectix though this was mostly used to dial in remotely to access the Windows systems. Looking back it was a great shame Microsoft bought and killed off Connectix.
RB builds for Windows at that time were notoriously buggy however with careful research and using the debugging tools in MS Visual Studio it was possible to identify what the issues were (mostly memory leaks consisting of orphaned objects). With careful programming it was possible to avoid these pitfalls in the Windows builds. The only problem I couldn’t solve was the memory leaks in NT itself, which had to be configured to reboot periodically.
Recently I was advised the system was finally decommissioned earlier this year - and they said they considered the apps as unbreakable - the system ran for 16 years without any further changes.
17 years later I can say the Xojo community has good reason to have faith in the ability of Xojo to produce robust cross-platform apps. I just hope Xojo continue to maintain their current product quality.
[quote=366785:@Nicholas Loveday]
RB builds for Windows at that time were notoriously buggy [/quote]
I’m not buying your story. ‘Compile for Windows’ they said, but I’m convinced that those early versions merely wrote a few MB of randomly generated bytes and then tagged ‘.exe’ on the resulting file name. 
Realbasic controlling a whole nation wide railway system? I don’t believe your story either. I am sorry.
I will not get on any of those trains on that railway. And doing that with only two people? Please be serious.
I think integrate data/information from several systems/sensors is not the same as controlling the whole system.
Not buying the story?
I remember this:
[quote=366821:@Jeff Tullin]Not buying the story?
I remember this:[/quote]
Now, THAT’s frightening. Anybody know where this guy ended up ?
A train controlling system is something completely different. Believe me, that is not designed and created by only two people. It is possible for a model railroad, however for a real railroad it is a very complex system, consisting of many sub systems.
RealBasic and Xojo are a very good developing language, however it is completely useless for a railroad controlling system.
Chris, maybe you misunderstood what Nicholas had to do. Or maybe I do. English is not my first language.
In you past post you said: “Realbasic controlling a whole nation wide railway system?” and what Nicholas said the needs were: “They wanted to integrate real-time GPS data from approx 2000 trains state wide, sensors that weighed and identified each vehicle, a billing system that delivered invoices in near-real time using SOAP, and deliver a graphical web-based user interface showing the path of each train that serves circa 10,000 internal and external users.”
I don’t see how, from what Nicholas wrote, you think he meant a “whole nation wide railway system”.
For me it translates to: “get GPS data, vehicle ID and weight, show the path in a graphic and a billing/invoices system”. I didn’t read that they will control the railway/vehicles.
Maybe because I worked as an integrator, I know that you can have a nice app with good graphics getting information from several systems and presenting that to the client. They may think that your program is controlling everything, but the “only” thing is doing is getting information from the systems they already have. Some systems will let you control them by sending information back, so there is really no need to create a new program to control the hardware directly. No need to create a full controlling system.
That’s what i read out of it too!
I am sorry, then I did misread and misunderstand. Also English in not my native language.