I often see calls for Autodesk to support AutoCAD on Linux. Just like AutoCAD for the Mac, while I can sympathise with the users of that OS, I think a native port of AutoCAD for Linux would be a bad idea. Again, I think it would be bad for everybody: Autodesk, AutoCAD for Windows users, and most of all, AutoCAD for Linux users.
Why? First of all, for most of the same reasons I gave for the Mac port. Autodesk hasn’t just failed in the past with AutoCAD for the Mac, it has failed with AutoCAD for Unix, too. I remember Autodesk being very enthusiastic about the Sparc port in particular (AIX, too). I know personally of customers who were caught up in that enthusiasm and invested heavily in a Unix environment, only to bitterly regret it a few years later when Autodesk abandoned them. Would this happen again? Probably.
Second, the numbers just don’t add up. Current PC OS market share is running something like this:
Windows 88%
Mac OS 10%
Linux 1%
While the Windows share is currently falling (thanks, Vista) and the others are steadily rising, there’s a long way to go before Linux has the numbers to make the investment worthwhile. In any case, it is likely that most Mac or Linux users of AutoCAD wouldn’t be new customers, simply existing users using a different OS. Not much of a cash cow, is it?
I dislike the Windows monopoly and support the open source movement, so I would love it if Autodesk could just snap its fingers and provide all its software on whatever platforms the users want. Mac? Sure. Linux? Great, why not? The reality is that it’s not that easy. It’s expensive to do and expensive to go on supporting in the long term. Unfortunately, it just doesn’t make commercial sense, and wishing it did will not make it so.
Does anyone have statistics on what OS that AEC firms use (or desire)?
When I was a CAD Manager, I would have moved every CAD Station to the Linux platform if Autodesk offered a Linux version. I suspect that those demographics would change…
The OS share is even worse (from the Mac/Linux point of view) in Europe or e.g. in Czechia.
MacOS is only at about 3% in Europe:
http://gs.statcounter.com/#os-eu-daily-20080701-20090610
and at a mere 0.99% in Czechia:
http://www.iinfo.cz/tiskove-centrum/tiskove-zpravy/navrcholu-operacni-systemy-08/
Linux is at about 1% in both regions.
The data is measured from access logs of thousands of web sites (general mix of web sites) – which is quite a significant method. Interesting are also the trends – Linux going down in the last year, MacOS unchanged or slightly up.
Here are the hits from this blog:
2008
Windows 91.9 %
Unknown 4.3 %
Macintosh 2.4 %
Linux 1.2 %
2009
Windows 90.6 %
Unknown 4.0 %
Macintosh 3.1 %
Linux 2.0 %
This is a biased sample, as this blog has a Windows-only product as its main topic.
As with the Mac, Linux poses the same chicken-before-the-egg argument. Do you run only apps for a particular favored OS, or do you pick your apps first and OS second?
Some people think that the source of all of their performance problems lie with Windows, as if Microsoft’s poor coding practices was the sole cause of all AutoCAD crashes, long load times, slow rendering rates and so on. Like their work would be 200% faster if it were just on another platform.
I think this is downright silly. These arguments are at least knee-jerk reflexive (I hate Windows, so anything else must be better), completely ill-informed (Man, those Macs are sexy beasts. Look, all the hip folks in the coffee shop run them. They must be better. Hm, just wish I knew how they worked), or outright disingenuous (AutoCAD on Linux would rock because it would use more efficient code!).
For people who want AutoCAD for , they should first answer these questions:
1. Do you have the necessary experience running high-powered apps on that other OS to qualitatively say you could work better/faster/smarter than on Windows?
2. Since the core of the application largely stays the same between OSes, any performance/stability benefits to the end user would mostly be due to the OS itself (more efficient execution, better memory handling, lower overhead, etc.) or in the compilation of the app for that OS.
In that sense, is Windows really THAT BAD? Seriously? Have you run the 64-bit version of AutoCAD on 64-bit Windows? Have you tried the beta for Windows 7? Being that modern Windows has a lot of its roots in Unix, it’s tough to see how something else could be so much better.
3. Do you really want to invest in all new applications for this OS? It’s a good bet that either you or your company has lots of money tied up in PC software licenses. Although there are lots of excellent free apps out there, there are some things which cannot be substituted for Open Source or low cost alternatives.
I cringe at the thought of what it would cost to move all of our Adobe licenses over to Mac or Linux. Oh, wait. Adobe doesn’t make Photoshop for Linux.
4. I argue that most of the problems with AutoCAD are the fault of Autodesk, not Microsoft. Do you really trust that Autodesk can pull off a Mac/Linux version that is at least as stable as the Windows one? Really?
And if it not clearly more stable and faster than the Windows version, what would be the point of switching?
5. Do all of your peripherals have Mac and/or Linux drivers? Are they stable?
The point is, moving to another platform for the sake of one application is at best a costly, time consuming and chancy effort.
How about running AutoCAD on AIR, from Adobe? AIR has different versions for Mac, Windows and Linux, I believe. That way Autodesk only has to run on AIR and they only have to devote time for one version and then you can use AutoCAD regardless of your OS. the concept makes sense, but what’s the reality?
That is a bad idea if you say moving with change in time for the goal of achieving the faster and reliable independent software in hand of common user, is bad. No one can waste their life more being dependent on either windows or Linux we all recommend is maximum service.
I remember the non-official reason that there were no further Unix versions of ACAD. Unix was moving ahead at a negative nine percent growth rate. Windows was the only stable platform with a multi-directional future and was the only responsible choice in the mid 90’s. I don’t see that much has changed. -Bill
Could someone please explain “modern Windows has a lot of its roots in Unix…”. Complying with the POSIX spec doesn’t mean that it has roots in Unix.
How about the knee jerk reaction, “I like Windows and have contempt for everything else…” My problem is the way Microsoft does business and the business expense related to the way they operate. Read your EULA and tell me who in the agreement is in breach of contract, the way I read it, it is Microsoft.
Answers to questions:
1. Yes. Users should also consider the investment cost of the OS and long term total cost of ownership. Don’t believe MS numbers, they tend to forget that viruses (and patches) on Windows actually use up man-hours and resources.
2. I have Windows 7, nicely done MS! Though it doesn’t mean I am ready to switch back to Windows. I will buy the upgrade and wait to see if the weekly patch model goes back into action. Microsoft can blame it on hackers and viruses, but wouldn’t it have been simpler not to do retarded things like activex and .net? It would be nice if I could just remove this junk from the operating system.
3. You assume that all of us use only Windows apps. The only Windows apps that we use are Autodesk apps.
4. Sure they can pull it off. They currently have cross platform applications with experience upgrading them. No, I do not believe that there will be a 200% increase in speed (as mention in a post). It is the same processors that we will be using. Why switch? To have the option of using a different OS with a smaller footprint, without the licensing structure, fees and total cost of ownership.
5. I haven’t run into any problems.
For the most part companies seem to be happy with Windows. No one questions about patches, service packs, viruses or activex. In our minds it is just the way it is. The costs related are just a cost of doing business. We wait for Microsoft to bless us with a new OS that will never fulfill the same promises they give at each upgrade. Do you ever stop and think, ‘there has to be a better way?’ If there is we’ll never get it if we are locked into one operating system.
This may change everything – Google’s new full blown OS – based on …… Linux !
http://googleblog.blogspot.com/2009/07/introducing-google-chrome-os.html
I saw that… and already the Microsoft fanboys are out saying ‘Why Chrome OS will fail’, ‘Google never follows through’, ‘Linux is hard to love’, etc, etc.
For scientific research I would like to a lot of modeling using AutoCAD but having to install Windows to do so made me search for alternatives first. Unfortunately I haven’t found anything yet which would suffice but this really kind of “grinds my gear”. How can customers *not* demand any kind of OS-choice?
Customers can demand what they like. I can walk into a Rolls Royce showroom and demand a pink one, and if I have enough money to throw away I may eventually get it. If I demand that it’s a hybrid turbodiesel too, then I may be out of luck. The development would be prohibitively expensive when divided among the small number of potential customers.
The problem is Autodesk made some bad decisions with the tool kits they used to develop the application since R14. That is largely why porting it is so costly. Had they used a cross-platform tool kit it would make sense. I’d argue that Autodesk should consider switching to platform neutral technology to avoid problems in the future. Doing it all at once as people abandon MS solutions is going to be allot more costly. While GNU/Linux might be 1% Mac’s market share has risen once again since there near death experience in the late 1990s and GNU/Linux has gained strong support from various companies. Canonical is owned by Mark Shuttleworth- a technically competent individual with the means to challenge Microsoft. Many companies are following his lead in adopting and/or supporting GNU/Linux. Since Ubuntu came to be Dell has (despite the crap support) done more than just test the waters. Other important thins have happened to. IBM, Sun, and others have developed a strong GNU/Linux portfolio. IBM rewrote Lotus Notes primarily because of GNU/Linux. It had no choice as they used the Microsoft tool kit to develop their solution. The point is companies are making the investment. It is just a slow process. We also have new start-up companies that are providing a solid catalogue of products and support aimed at the non-technical audience. ThinkPenguin now sells a competitive line of systems comparable to Mac OS X in quality and has the products/accessories/support to make it happen for most non-technical users. You can actually buy a system that has full support for 3D acceleration, suspend to ram, wifi, and all the other chipsets- plus based on free and open source drivers-ensuring compatibility with Ubuntu updates. In comparison Dell released systems with non-free drivers and provides zero guarantee of being upgradable. Trying to use a distribution from 2008 on your laptop in order to get full use of it just doesn’t cut it. And in fact you don’t get full use of Dell’s laptops since non-free graphics drivers prevent things like suspend to ram from working or working properly.
My point is that GNU/Linux is gaining the support non-technical users need for the OS to take off. We will soon have a source for GNU/Linux solutions at the local level- similar to the Apple store. If I had to guess how far off we are to that happening I’d say a year. Freegeek- a non-profit has shown it works. Their angle is not really GNU/Linux- they just utilize it to “help the needy get nerdy”. So the next step is to commercialize that type of operation and open stores where users can get support. Obviously employees will replace volunteers and new products will be offered rather than recycling. What will happen from that point is the end-user dollars from support/sales from these stores will go to developing new open source software.
For your information: Bricscad, which is fully compatible with AutoCAD’s DWG format, will release a native Linux version within 3-4 months fron now. There is already an alpha version which can be downloaded to get a flavour of how this product will look like. See https://www.bricsys.com/common/support/forumthread.jsp?id=11642 to get the download link for this alpha version and to follow the forum discussion about this Bricscad Linux version.
If one still remembers the nineties, there was a lot of discussion: which OS would be dominant, OS2 or Win95. One IT journalist, Dvorak I think, was very adamant: Win 95 because long before the launching of the OS, Microsoft was encouraging the software developers to create Win95 versions of their software. And he was right: the OS with more software written for it has the advantage and the appeal. Now, OS2 is not even a memory in most of us – it has completely vanished – mostly due to lack of specialized software. I’d love to use Linux, but is not ripe yet – OpenOffice, Firefox, VLC and some other general-use software is not enough. The meat is missing from the plate. Maybe later…
All the big EDA vendors, like Synopsys, Cadence Design Systems, Mentor Graphics etc. have always offered their software natively on Linux/Unix. That’s why I find it weird that AutoDesk is MS-only. Of course EDA software can be way more expensive (some Cadence’s software cost $100000 – $200000).
The numbers for the operating systems dont tell the whole picture.
Windows is dominant in the PC realm.
PC stands for Personal Computer, by the way, as in the cheapo-crapo at best buy.
The workstation market, linux dominates with mac coming in second.
For servers, its pretty much a photo negative the PC market, with linux sharply dominant, near 90%…
Business Users of Linux
Novell
Google
IBM
Panasonic
Virgin America
Cisco
ConocoPhillips
Omaha Steaks
Amazon
Peugeot
Wikipedia
New York Stock Exchange
Burlington Coat Factory
Raymour and Flannigan
Tommy Hilfiger
Toyota Motor Sales
Travelocity
Government Users of Linux
U.S. Department of Defense
U.S. Navy Submarine Fleet
The City of Munich, Germany
Spain
Federal Aviation Administration
French Parliament
State-Owned Industrial and Commercial Bank of China
Pakistani Schools & Colleges
Cuba
Macedonia’s Ministry of Education and Science
U.S. Postal Service
U.S. Federal Courts
Government of Mexico City
Garden Grove, California
Largo, Florida
Czech Post
Educational Users of Linux
Russian Schools
German Universities
The Phillipines
Georgia
The Indian State of Tamil Nadu
Switzerland Schools
Bolzano, Italy
Kerela, India
One Laptop Per Child
Indiana Schools
Home & Scientific Uses of Linux
Sony Playstation 3
Netbooks
Some Dell Models
CERN
Internet Archive
ASV Roboat
IBM iDataPlex in Canada
HMMMMMM…………AutoCAD costs enough, I think that they can use a little of that revenue to explore other OS.
Retarded.
Everybody knows now (2012) that linux is the present & future.
Windows is quickly becoming the past.
Most new ideas, applications, etc. from Windows, are just regurgitated from what linux has already accomplished.
I laughed when I 1st saw the Windows7 commercials, bloating what it could do.
LOL.
My old linux system already did all that, with 2 year old firmware.
Fanboys are cute.
AutoCad fanboy speaks irony.
It’s ironic that anyone would consider me an AutoCAD fanboy, given the contents of this blog. I’m sure the Autodesk people who cringe every time I post something would be highly amused by that.
Second, the numbers just don’t add up. Current PC OS market share is running something like this:
Windows 88%
Mac OS 10%
Linux 1%
Market share?? LINUX IS FREE!!!! Windows is almost always included with a pc, i have at least 4 licenses, so in the market share i bought windows 4 TIMES! but I do not use windows, first thing I do is formatting the hdd and installing a free linux becous its better and its the future. for instance, everybody is beginning to use Android… ANDROID = LINUX.
So is it a good idea to port your software to linux, yes it is!
Statistics: the art of torturing numbers until they confess.
Time will prove that this blog is wrong.
The major problem is that the author of this blog cited statistics of OS usage as a whole (business home and otherwise), this is crazy. Autocad and friends are BUSINESS software. You need to look at the OS usage stats for businesses. When you water this down with home computers, you get the wrong results.
If you look at the industrial graphics and big data industries (movies, animations, TV news, newspapers, accounting, banking, medical care…), you find that they are completely dominated by Linux with the one-off exception of CAD… Why is CAD the exception, and why did the author of this blog turn a blind eye to this fact?! These two subjects would make a great blog, this is what we want to see.
Actually, time so far proofs this blog is right. BricsCAD for Linux is almost 7 years old and what CEO of Bricsys says in a recent interview:
“Still, we must say that applications availability for Mac and Linux is not much. That has to do with 95% of our sales being on Windows. We expect that might change for BIM, because more architects are Mac users, partly because the first version of ArchiCAD was on Mac, Vectorworks is, so it’s really an Architect’s machine. We expect that maybe for BIM, it might change and we might sell more versions on the Mac.”
It is not just CAD alone – it’s a whole plethora or businesses apps around CAD that are just Windows only.
Funny, I was just thinking about how time has so proven this blog wrong, we are now in an age where what platform you are using is becoming less and less relevant due to online versions of things like Microsoft Office and how there is now so many more options for those running Mac and Linux than there used to be. Even Autodesk is making AutoCAD for Mac, yes it still needs some work to bring over some missing features compared to Windows, but it is far more stable than the Windows version and runs faster. There is also BricsCAD and Draftsight that have Windows, Linux and Mac versions that all work identically no matter the OS.
Overall we are now at a point where we can easily choose our platform and have compatibility.
This post is specifically about Autodesk creating AutoCAD for Linux, and that being a bad idea. It remains 100% correct. There isn’t an AutoCAD for Linux 8 years later, and there almost certainly won’t ever be one, because it doesn’t make commercial sense.
Even the Mac version never justified its investment, which is why it’s been left half-finished all these years and hasn’t been followed up with Mac versions of other product lines. Even AutoCAD-based product lines.
I’m all for people running Linux if they want to. I have a couple of Linux PCs myself. Browser-based stuff can of course run in anything, if you’re happy working that way with its fundamental limitations. Great.
BricsCAD for Linux, also great. Because of the underlying architecture of BricsCAD, Bricsys gets that and the Mac version almost for nothing. I’m happy that Linux users have that level of choice.
Autodesk wouldn’t get AutoCAD for Linux almost for nothing. It’s way too hard to do, given the likely return on investment. That is, it’s a bad idea.
I dunno. Dumb ideas with a really bad return on investment seems to be right in Autodesk’s wheelhouse these days. 😀
lol