As a retired EE, I do some design for fun, and some design and analysis work on a contract basis. For some of these tasks, I need Word, Excel, and PowerPoint (I've tried the Open Office equivalent tools, and there are problems with these tools for my needs). I assume that for both tablet and desktop environments MS will continue to offer these tools. For other tasks, I need some form of schematic capture, PCB layout, and simulation tools, as well as math tools (MatLab, MathCAD, for example). All of the tools that I use presently are Windows based, though there are some Linux based work-alike tools. For my needs, tablet versions of any tool is simply not an option. I also do not see cloud based versions of the tools as viable options for various reasons.
During my working career, the majority of the serious design tools were Unix based (and far out of reach for home use financially), though toward the end of my career our tool vendors moved toward PC/Windows based tools. The PC/Windows versions of the tools were a giant step backwards in capability, in my opinion. With those tools we were able to produce schematic drawings and board layouts just like the ones done by hobbyist tools, note sarcasm. Hopefully the serious design tool vendors will migrate their products to whatever Windows version that emerges, or will offer Unix/Linux versions as they once did.
I came across a Red Hat suite, Fedora Electronic Lab (link: http://spins.fedoraproject.org/fel/) that purports to provide an electronic design suite suitable for board and VLSI design, though I haven't tried the suite yet.
After this diatribe, I guess my hope is that MS continues to support desktop versions of the Office suite, and that the tool vendors offer desktop versions of their design and analysis tools compatible with whatever O/S Microsoft offers. Alternatively, I hope those vendors offer serious versions of their tools that run under Linux.
My concern is that if Microsoft coninues to push towards the tablet only interface and ignores the development interface presented currently under Windows 7. Previously Microsoft had embraced two separate versions of Windows when it went off with Windows NT (I think it was), but they seem to be studiously covering both bases with WIndows 8 RT and Windows 8 Pro. Maybe I am wrong- perhaps RT will be the way of the tablet as the number of apps grows and Windows 8 Pro will return to (or at least stay with) its desktop roots
Win8 actually does work for development. I've been doing it since the preview release, so had to deal with the blue screen of death periodically for the first few months. But that's been solved by further updates and now I just press Win-D after logging in to get to the old familar Desktop.
There are annoying things about the interface. But the driver situation is even better than with Win7 -- most drivers appear without requiring installaltion disks (in fact, I haven't needed one). And with a boot SSD and software RAID array, it works the same as it did under Windows 7 -- all development tools and previous Windows apps installed just fine, including development tools, IAR, and Code Composer.
it is all about the interface. That old tried and true windows desktop falls apart when placed beneath a touchscreen. We know this, we've tried it for years! The new metro interface is a step in the right direction. Unfortunately, without a touch screen it feels extremely foreign and uncomfortable. This leaves the current windows in a state of dual personality. Are you going to work as a touch screen environment or go to the old interface?
I think the death of the PC as commonly described in the popular and industry press is premature. Sure, PC sales have fallen. I think part of this is due to no pressing need (and even resistance) to upgrade to Windows 8. Part is also due to no great need to upgrade the PC for performance reasons. Certainly a several year old PC can easily handle typical email, word processing and spreadsheet applications. I don't need to do a lot of extensive simulation or other CPU intensive engineering so am able to do my development and CAD work without a top line PC. Instead of a new PC upgrade on my usual cycle I bought a tablet. I bet a lot of others have done the same.
I really like my tablet, but to do real work I need a more powerful machine, a keyboard and a couple of monitors. I can't imagine doing a CAD drawing on a tablet. This is not limited to the engineering community. Although I see a lot of tablet and smart phones used around here, I can't imagine our accounting, marketing, graphics and other departments giving up their PCs for tablets long term. Business users will continue to buy PCs.
PC sales might have peaked, but sales will not fall off a cliff. We should learn a lot next spring. XP will not be supported after next April. This means there will not be any more security patches and this will force many companies to upgrade. I have heard as many as 70% of all PCs in use run XP.
Business users are going to be a large market for a long time. I don't see them moving away from Windows although they might skip some versions like Vista or Win 8. Most engineers will be given Windows machines for their business needs and expected to run their engineering applications on the same box.
Developers of engineering software will be forced to move their products to newer versions of Windows.
Hi Betajet, I believe there is already a Xilinx port to Linux. I personally welcome this move, it means I don't have to spend megabucks on SW every time MS creates an incompatibility. Currently I'm running XP on my development machine because my $22,000 PCB and simulation package won't run under Win7 (and certainly not Win8) so I'm keeping it going by hook or by crook.
They used make sure that older SW always ran, but around Vista they decided there was more money in pissing off customers so they wouldn't upgrade their PC's. I personally think that the the dent in the PC market is because it's no longer $1,000 for a new PC, but anywhere up to a $50,000 per PC for some companies.
I know companies that run their accounts on SW that runs under DOS still, and they need (or want) complete history so they don't want to change that accounting SW and keep old XP machines running at all costs.
After having dipped my toes in the Linus water, I have to say that I cannot agree with this statement. I do realize that Windows does tend to make major changes, but I have to say, I rarely have run into difficulties getting software to work across multiple versions of windows. Some software that was designed to only work with Win XP has worked fantastically with Win7. That is a pretty long lifetime in the computer world to have an application that is 10+ years old and still transfer across that many operating systems.
In contrast, while I was using Ubuntu, I stopped trying to upgrade Ubuntu because every instance I did, I would break the installation of not just one, but multiple programs. I would then spend a few days trying to get these working again. Having to deal with this every 6 months would be untenable, though not applying updates could leave you prone to not having features that are time savers as well as security improvements.
The reason I stick with Windows is that all the programs I need work with little effort. Linux has come a long way, but I do not see it as a viable alternative for the masses in most of its current forms. Even Android suffers from the fragmentation problem, and it has not even been around as long as Win XP.
In a recent Dell advert, I noticed that all the PCs intended for business came with Windows 7, and all the PCs intended for home use came with Windows 8. I think most businesses are planning to skip Windows 8, just like they skipped the last even-numbered Windows version (Vista).
Another attractive solution for business is to run virtual machines on servers, and have thin clients on the desks -- or on tablets. Centralizing computing in this way reduces maintenance headaches. Again, back to the future: this is not unlike an IBM mainframe with 3270 terminals on the desktops except that you can move the equipment without getting a hernia.
I'm most familiar with Xilinx design software. While it's generally most convenient to use the GUI, most of the work is done at the command line interface level. For the most part, the GUI just generates files and commands for the CLI and then extracts results from output files to display on the GUI's summary page. Xilinx has a good CLI reference manual, so you can write Make files that automatically run the tools after a design change, which is a better strategy for product build.
In Xilinx's case, the CLI tools never really left Unix. I believe they always had a Unix version (usually for Sun) so porting it to GNU must have been pretty simple.
Some design software is Windows-only and some is integrated into the .NET framework. A good example of this is the PSoC Creator tool from Cypress, which I suspect would be a nightmare to port to GNU. Design tools pretty much standardized on Windows in the "aughties" (200x), but in that time GNU/Linux progressed remarkably and nowadays the only reason to run Windows is for legacy software. Engineers are starting to see that GNU/Linux is a better and safer platform for Serious Computing, and companies that insist on Windows-only risk losing sales to competitors who offer choice.