One of the problems effecting Unix/Linux in their contest against Windows has been the fact that first Unix and then Linux have been houses divided against the concerted and well researched attacks of Microsoft marketers. Too often to recall without grimacing, Unix developers have rallied around the flag only to dissipate the rally in a muddle. Give Redmond credit. They make it their business to know the strengths, but even more critically, the weaknesses of their opposing software. And as it turns out, Running Linux by Matthias Dahlheimer and Matt Welsh, is a perfect document to underline the strengths and weak points of Linux. So this review will be done from the point of view of a Microsoft Windows Product Manager, Stefanie Kugelsee, and how she might be assessing Windows versus Linux distributions. Here is how Stefanie's analysis might proceed:

Linux Overview Chapter

Typical Linux approach - this is a 1000 page book on Linux and the authors barely scratch the surface on what makes Linux desirable. And there is no direct comparison with Windows or any other OS for that matter. Our Get the Facts website appears therefore as a treasure trove of facts in comparison with the various Linux magazines, websites and blogs. And of course Slashdot.com and its Linux-blog ilk are like liberal politicians in the US - mammoth flak catchers diverting attention from us with their often rude and insensitive commentary on the computing scene.

Nonetheless the Linux Running authors manage to scratch the surface on 4 areas of concern to us in Redmond. First, the proliferation of the Linux desktop in China, India, South America and even parts of Europe is being addressed with the lite-price version of Windows XP and our own continued ambivalence about Windows and Office piracy. An amnesty with payments and tax credits may yet allow us to bring Chinese and other delinquents into the revenue stream fold. Validation and subscriptions are keys to "bringing piraters on board".

Second, the authors barely touch on Windows bloat. Linux can run with good response time in 1/2 the memory space and 1/2 the CPU speed of Windows Vista. I told Allchin we would need some performance tuning this time around - and then the security bombs hit worse than 9/11. Who could have connected the dots ? The problem is that with all the publicity on dual boots with Apple and virtual machines, inevitably somebody is going to do the tests comparing Vista with Linux and Apple and Jim's going to be low man on the performance totem poll.

And now we have Windows Rot and Clot to contend with. I told Jim we would have to clean up the registry and our configuration woes. Now Windows Rot means Windows OS performance has a permanent degradation cycle reversible only by a re-install. Ditto for Windows Clot - this time malaware and viruses so corrupt the boot area, registry and other system configs - that complete wipe-disk and re-install is required. I warned Jim - but he is like Donald Rumsfeld - hanging on and in effect not allowing us to address these problems..

The third headache, the continued ascendancy of LAMP in the Web world is at least heading in our direction. Brilliant ploy to give Godaddy a sweet deal for conversion to Windows Server 2003. With all their parked, unused webnames - Windows Web server count went up by 6% while LAMP registrations declined by 6%. But really, on the Server side we finally, after 12 years, got the availability and reliability numbers of Server 2003 into Linux range. As I always say "in case of near miss or tie, we win".

But we can put LAMP further on the ropes. PHP is starting to get mainstream support from Oracle and IBM but AJAX is the new development darling. And we have our one Atlas solution tied into Visual Studio. Meanwhile Running Linux authors don't even mention AJAX yet there are over 10 AJAX versions coming from Open Source. Heh heh ... it looks like a microcosm of what we are doing to Linux in general. They have hundreds of distributions of which 7-12 are major players. But each of these players has its own install, its own set of apps, one of three GUI frameworks, one of three major office environs, one of 7-8 shell command languages - no wonder we can claim that Linux system administration costs will be as high as double those of Windows(good thing security is not a part of system admin).

And now that Bill has said "mea culpa" and has gotten the IE7 team working hardcore on cutting off any more Firefox erosion - we should be able to establish our Web hegemony once again. Look at the coverage of Web development in this Linux book(see below), it is minimal. So we will determine, not W3C nor WSI nor anybody else, what Web standards actually get adopted. Sayonara to XForms, SVG, XPath2, and any meaningful changes to DOM. Boy the Mea Culpa ploy was a brilliant use of minimal concessions to standards promised 7 years ago - and to get them hooked on yearly installments of IE7. Bill-Brilliant!

Fourth and finally, Linux has the same problem we do they just don't seem to want to recognize it. Desktop computing has split into many specialized working sets - call them Web terminal users, desktop dummies, designers, developers, analytic users and power seekers. I am still not convinced our Vista line up is appropriately tuned either. But if we are off base, Linux Running is off lost in the desert - with hardly any attempt to classify general PC user groupings and needs. And likewise there is little effort to match those needs to what the various Linux distributions are offering . But equally unforgivable - Linux vendors barely mention the other critical advantages they have over Windows. If Linux was not free it would be a lost cause which we would not have to worry about.

Install and Getting Started Chapters

Why would anyone want to read about a generic Linux install that does not apply in detail to any specific Linux version? The whole problem with Linux is that they have not only unique sets of features but also the install program and the install file formats vary from here to eternity among distributions. Okay I exaggerate; but see here and here for a comparison of ease of installs between Windows and Linux. And we still have better support for more hardware devices and peripherals than any of the Linux distributions. The authors confirm our advantage when they start talking troubleshooting IRQs and DMA channels. So for both crucial installation and system administration - we beat them hands down.

And look at the KDE, Gnome and Linux terminal chapters - all big winners for us. KDE and Gnome start getting close; but then we put in semi-transparent menus, 3D workspaces and viewers, plus shuffle ports - and they have to play catch up yet again. Linux terminal mode might be a worry if someone developed a character mode GUI - it would be lightening fast. But the confusion of working in terminal mode prevents that. There are so many different and cryptic commands+options, the variations among shells and versions plus constant variation in the distros as to where apps can be found in which directories and named in what fashion - Linux is making an Abbot and Costello mockery of itself. And as for man versus our Help system - bring them on. In short, Dahlheimer and Welsh provide us with ample ammunition here for discrediting Linux..

As for Web browsing and mail services - we should have our shop righted - better IE7, Outlook and Exchange services linked to a fundamentally better WinCF. I concede FireFox/Konqueror plus Evolution mail may give them some temporary advantages. If mail goes upto the Web as Software Service our Hotmail will beat Yahoo and Google flat out in a couple of years. But the book's coverage of public key encryption of mail shows why our security tied email - Live or desktop, is a winning combination - so much simpler.

When Running Linux moves out of the generic install and into application selections it shows how far Linux has come - and also how we have covered off any possible openings. Lets just compare application offerings. Sure they have many more for free; but Windows has the most by far and our winning formula - the most applications at the state of the art including a growing stable of more tightly interconnected Microsoft apps just carries the day - look what happened at Bank of Canada.

Chapters on Application Offerings

Nowhere is my thesis that we do more applications, and better, illustrated as well as with GnuCash. The authors spend 17 pages covering GnuCash. And GnuCash is impressive with some of the dialogs supporting cash accounts worth the time for our developers taking a closer look. But face it, Quicken and especially our Money are a step beyond with routines for loans, personal finance, an easy step up to full business accounting, taxes - the whole home-business financial services. And we find its our reports and tie-ins with various financial institutions and financial information or advisory services that keeps people coming back. You just can't give this stuff away for free - you have to reward some of the major financial players with something of value - and the Linux guys and gals are anti-profit just killing acceptance in the "quid pro quo" world of financial services.

Multimedia is another example. As the authors point out, Linux is making inroads in this sector - especially with render engines for 3D plus high end video and other multimedia apps. That May 23rd 2006 issue of PCMagazine issue devoted to HiTech Hollywood is a disaster - at least we got them not to say "How the Linux PC saved Hollywood". But just like the turmoil in our graphics file support (why cant we just get these PNG and JPEG2000 issues solved?); the Linux guys and gals have sound plus graphics support problems with ALSA, OSS, and rival commercial sound drivers making sound a scratch proposition in Linux. And look at the number of choices for multimedia toolkits: SDL, OpenGL, NMM, MAS, GStreamer, etc. Adding FUD to this multimedia soup should be easy.

And again, the authors spend nearly 20 pages going over the advantages of GIMP as a substitute for pricey Photoshop - and I have to agree, its not half bad. But our Expression Suite with its close ties to XAML which super powers XUL+SVG will win the day. Expression alone links bitmap and vector graphics directly into 3D and animations. We are now a step ahead of even Adobe+Macromedia. And after GIMP and the highend movie stuff, Linux is a wasteland for good media software. And there is just no distro that even comes close to matching our Media Center Windows and all its media savvy apps.

The coverage on Games and Groupware is typical of this book and Linux in general - so close and yet so far. Everytime they get close we move away. So what if their cards and arcade game line-up beats ours. And even if their games now can do 3D and networked action games - our new Vista Presentation system solves a lot of Windows and Web and 3D and High Definition media problems that Linux is just starting to address on the consumer end. I know, I know - PCMagazine's May 23 issue is devoted to Linux highend media- and how all the big Hollywood shops have gone to Linux on their render-engines, servers and workstation desktops. Industrial Light and Magic, Dreamworks, even WETA Digital. But this is all highend (read $$$,$$$) stuff - nobody has brought Linux into the desktop media fray. So Linux is to us in desktop GUI and media as we are to Apple - constantly playing catch up. But hopefully by taking the hardware bloat hit, we step ahead of Apple.

Ditto for Office. So they have Open Office, Koffice, and Think Office or whatever - our new Office 11 for Vista is going to offer so many new services and tie-ins that we will just expand our share. Yet these authors dare to raise the Open Document Format issue - we own 90%++ of the Office market so we should set the XML storage format standard. Macht ist Recht.

Open Source means free access to software but not totally free software. This means software that is made available through Open Source license has made its source code plus resulting executables freely available to its users and, depending on the Open Source license used. But distribution, support, training and other services are just as expensive as our own or other commercial software.

But here is the inside goods. Open Source is certainly a reaction to SaaL-Software as a License vendors like ourselves who charge an up front license fee for their software plus also charge for distribution, support, and training. EULA and other licensing terms are anti-warranties. EULA and other SaaL licensing terms are constructed to deny the software buyer any warranty or legal recourse in the event that software fails to work as advertised. Unless we improve our reliability, security, interoperability to the point they are good as the others- and keep the marketing Yahoos off their Exorbitant Promises HighHorses - the EULA is going to kill us.

I keep telling Bill and the guys even with all those apps and all the sunk training costs, that won't be worth a hill of beans if we are the highest price producer and the perceived value of our service and support does not match the Open Sourcers. This will be especially true with new Vista and Office/Vista editions because both programs have big retraining costs.

System Administration Chapters

I love reading the System Administration chapters of Running Linux. It is all about mounting files and filesystems, use of cryptic shell commands and even recompiling the kernel. Now this may be fine for some server side administration types but for the desktop user, this is all mortal sin territory. Even hint at that stuff to the average consumer -and Linux will be dead on sight. No need to use FUD. Just point consumers to the appropriate pages in Running Linux or the Suse Linux or Xandros manuals(ahh fond memories, reminds me of Comdex 15 or so years ago when we almost single handedly took out OS/2).And every Linux distro has at least two or three chapters in their docs. Running Linux has 10 chapters and 300++ pages on the pleasures of "losetup -e +vv -cipher /dev/loop0/ FileDHere. And the poor Linux bastards - get one letter off - uppercase when it should be lowercase or vice versa, and their command script is dead in the water. I love it! Got to make sure everybody knows this crippler far and wide.

But here too, we have to be careful. We have our own WSH + VBScript + Perl+ VBA +GKWE scripting mess. Fortunately MMC consoles puts all this into dialogs and under GUI control. But Novell's Suse Linux will be adding full ZenWorks - and that could be real trouble. I have told the guys we have too many different and scattered admin consoles - lets not do a Unix/Linux here!

But really - system administration is becoming the critical point in sales of the desktop in many large organizations. Look at how SaaS-Software as a Service vendors are billing themselves. They reduce your networking and hardware infrastructure requirements - and reduce system admin down to zero. Now if SalesForce.com and buddies just could stay up long enough to not incur huge SLAC-Service Level Agreement Cost penalties, they might really take-off. We better get Live quickly. Note to self: find out whose Server OS is providing such wondrous service to the SaaS gals and guys. Also, who is best at running huge clustered service farms?

And this whole System Admin thing is behind the rise of computing's Living Dead - the zombie idea of thin client stations. We and Citrix have beat them back too many times to hell. God I am glad we left those old OS/2 Citrix guys alive. But here is Thin Clients reborn not once, not twice but three times. We think the old 80-20 nonsense and system admin costings will be out weighed by the sheer convenience of having a fully capable client.

Fortunately, Running Linux authors devote two full chapter to more living dead software in the chapters on Text Editing and Processing. At least we have the intelligence to keep scripting and our own dork zombies in the back pages of MSDN magazine. But 50 pages devoted to vi, emacs, regular expressions, LaTex and other Linux software zombies - sweet. After thin clients indigestion, it cheers my heart to see the Slashdotters shoot themselves in their feet with antediluvian software. Hell, we cheer every Linux distro that excludes Eclipse and NetBeans - now these are real editors and competitive IDES. Of course, not a word in Running Linux on either tool - God, I love these authors. So you can imagine my enthusiasm for the chapters on programming where there is more of the same.

Programming and Development Chapters

Programming has become like car design or movie making - a specialist enterprise. It is rocket science now that n-tier heterogeneous systems require expertise across so many arcane disciplines and languages (C/C++, C#, XML, SQL, HTML, WSDL, JavaScript, etc etc). There are so many trade-offs among GUI presentation and interaction, transaction processing and persistence strategy, workflow and asynchronous versus synchronous processing, and the constant problems of security, reliability and performance tuning. So authors have a pretty daunting task to get their readers up to speed in all the tools and services that are available in Windows or even Linux.

Fortunately for us, Linux Running Authors stick to the old guard, command line tools like CVS, gcc, editing Apache .conf files - not hinting at the wealth of tools available in GUI and now Web interfaces for linux development work. We have enough problems coming to terms with cross platform tools available from Trolltech, Java, Python and others. And we are still in denial about the impact of AJAX, Flash, PHP Perl, Ruby and other tools on the Web interface. But reading through Running Linux, the authors are still deferring to the command line gods.

I can't be bothered with the rest of the chapters because they look like mini-books within the book. Instead of giving developers executive summaries and resource links to the key and latest tools for Linux Development on a wide and broad spectrum, the authors have chosen to go deep on a few tools like PHP/MySQL and ProFTPD. Our win, Linux loss.

In sum, a book like this especially for the desktop consumers and workgroup managers, does a better job of striking fear into their hearts than any of our Get the Facts campaigns. I have got to persuade those marketers to just quote and link to the Linux gurus like these - and we will coming up winners everytime.

Summary

Who am I to contradict Stefanie or put words in her mouth. Rather, it is worthwhile to note the timing on this book. It originated in 1995 when a comprehensive view of Linux and its facilities were a rare commodity. But the authors appear to pay homage and devote a lot of ink space to command line technologies that should be reserved for the specialist books that are now available in the Linux market. Even as a guide to server side Linux administration the book is too brief on topic such as security, interoperability with other OS and overall manageability within large networks. Also the tremendous range of operating environs like embedded systems, mobile devices, and specialized servers (warehouses, management, clustering farms, etc) gets only cursory attention. This book needs to move up to an executive briefing level given the comprehensive and pervasive nature of Linux usage today - become a briefing and reference guide to all things Linux like it effectively was 10 years ago. Instead it still serves a developer/administrator audience that are finally now being well supplied with good books and other references like the RedHat and Suse Linux books reviewed here.

(c)JBSurveyer 2006