I read in a previous post that businesses will be more eager to use Linux since it does cost almost zero. I would like to point out some caveats that make businessmen reluctant to include Linux based Hardware and Software in their information system. Please note that by businesses I mean all businesses not just the editing software industry.

- Eventhough Linux and Unix have very similar capabilities, Linux is lacking *official* support. The main concern to businesses is maintenance. Contrary to what many believe, industrial software often have a lifetime of 7 years and more. If I experience any problem under Linux during that 7-10 year period, I want to be sure to have a support center that does know of my configuration and of all the Linux evolutions made through maintenance during that period of time. I do not know yet how much time I can rely on a Linux User Group: will it still exist in five years ? If not, what can I do ?

- official support strike two: the trouble with maintenance is that you need warranties of good services. If as a result of bad maintenance your software does not run anymore, what happens to your company ? In my line of business (financial markets), time is money: the cost of a non functionning software tallies around 10 million USD per hour (loss of financing capabilities, unavailability of payment systems, fines for not respecting contracts, ...). The use of backup systems does reduce that cost, but the residual cost of non functionning software is still important enough to eat up the income of the company. Thus businesses prefer to rely on official support systems with negociated service contracts: there are penalties for lateness in maintenance and for errors of maintenance. This means that *unless* you do have a professional support system able to engage itself on an important service contract (read multimillion USD penalties in case of failure of service), Linux will be marginally used in businesses: no one will want to base central information systems on that OS.

- Linux based software editing companies: well for the moment there are many companies developing software but they do not last. This is a true story: we bought from a good company (we had made financial analysis on the health of the company, the company existed since 3 years) a really good piece of software that managed Internet and internal company network communication (it was for a B2C type of service). It was working great when suddenly, the company collapsed (due to a theft from their financial manager). What do you do ? you have a working B2C service but no real backup and nobody anymore for the maintenance. This is a risky position where our business image can be harmed by a simple malfunction of that piece of software. We had to find out a proved solution to replace that software (that means UNIX based with big multinational companies). Of course, you may tell that nobody could prevent that kind of problem. You would be right, but what I did not tell you is that this company was the only one satisfying our criterias in the Linux world of business companies.

- Linux looks like a fad to business manager: the anti Microsoft hegemony is in, let's change all our systems for Linux. Managers would like to lower their cost but Linux does not benefit yet of a strong support system and nobody can yet predict if it will last against the heavies like Microsoft Windows and UNIX. Do not forget that to translate a great information system from one OS to another takes years (accounting system, database system, software managers, OS network intercommunication, and so on ...). A simple example: we had a software engineer who lived by Linux only. He succeeded to create a small information system in Linux for demonstration and testing purpose (the real implemented one was UNIX based). His own computer was under Linux. Two years later, he changed of job. The trouble was that we had his Linux based system but our UNIX engineers did not know what to do with it: we could not spare one person to manage that system. Moreover the specs of that system were written on Star Office and nobody else had that soft on its computer. It took us 6 months to retrieve the docs and to convert the Linux system to a functionning UNIX based one.

To conclude, Linux is still too young to be used by companies into their information system. Still managers are observing the market, and the fact that IBM communicates on the benefits of Linux on their mainframes (basket ball commercial) does interest a lot of people.

Ghostly yours,
Red.

[edited by - Red Ghost on May 6, 2002 12:25:19 PM]

Sorry to jump in this late... I think there are just 2 issues really, although they're big ones. The first is that Linux just isn't 'there'. Why develop games for a tiny audience? Even if you're just doing it for fun, it's not much good if nobody can play it. The second is that the OS is not all that compatible with games. Microsoft were willing to compromise stability and allow direct hardware access with DirectX, in order to lure game developers to the OS. Before that, they were sticking with DOS because Windows 95 just wasn't fast enough. What is happening with Linux to address this? OpenGL is ok for some things but is hardly as easy to use as DirectDraw was. SDL is simply not as fast as DirectDraw for 2D, even though it uses DirectDraw under the surface. Not to mention audio latency... Win98 is supposedly the best OS for this, while Linux (real-time kernel excluded) and later versions of Windows are pretty slow.

I think Linux gaming will take off if and when Linux on the desktop is common. This in turn is gonna require a combination of more and better applications for Linux, and the relevant action taken against Microsoft to ensure they can't muscle Linux out by way of bribing/blackmailing hardware suppliers.

[ MSVC Fixes | STL | SDL | Game AI | Sockets | C++ Faq Lite | Boost | Asking Questions | Organising code files ]

[edited by - Kylotan on May 6, 2002 3:39:18 PM]

quote: What is happening with Linux to address this?

There''s a couple of things. First is svgalib, which most linux users don''t like because it requires suid, but has been successfully implemented for games like Doom. The second possibility is framebuffer. Not entirely sure how this works, but I suppose it''s just direct hardware access. It allows you to run programs without X that would normally require it.(gtk+ apps, and IIRC, SDL runs on framebuffer, too)

For those of you who think there''s no official support or liability for linux, that''s what RedHat uses for its major revenue source. Businesses buy support and warranties from Redhat for a certain number of computers. IBM has a huge contract with Redhat, last I heard. Redhat would support the software IBM distributes on its hardware.

So "official" support is not really an issue. The issue, though, is that because of the range of possibilities(no support, amateur(IRC, LUG), consultant, service corporation), people often get mixed messages about the availablitiy of support.

quote: Original post by Red Ghost
- Eventhough Linux and Unix have very similar capabilities, Linux is lacking *official* support. The main concern to businesses is maintenance. Contrary to what many believe, industrial software often have a lifetime of 7 years and more. If I experience any problem under Linux during that 7-10 year period, I want to be sure to have a support center that does know of my configuration and of all the Linux evolutions made through maintenance during that period of time. I do not know yet how much time I can rely on a Linux User Group: will it still exist in five years ? If not, what can I do ?

IBM has shown some support for Linux, and Caldera owns UNIX (they bought out SCO), so those are two good heavies with whom to contract support. RedHat is another possibility, but to be honest the future of that company is unclear. I''m not sure if LinuxCare still exists, and I know that VA Linux folded, so you do have a point. However, I''m banking on Big Blue to become a major Linux player.

Sun is also toying with a full switch to Linux, leaving them to develop hardware and market support. Currently they''re using Linux as a migration platform (offering it on their lower-end systems) to their more advanced Solaris machines.

I think that also answers the warranties question.

quote:
- Linux based software editing companies: well for the moment there are many companies developing software but they do not last.

I advocate the strategy of allowing individuals to develop software independently and then have corporations take over the support and management of those products (RedHat and Bugzilla, GNOME to a certain extent). The formation of the GNOME Foundation and the KDE League also bodes well for those products; if we could have more similar collaborations between commercial entities that stand to gain from the software, their futures could be better guaranteed.

quote:
- Linux looks like a fad to business manager...

This is the fault of many Linux advocates (and the reason why I strongly dislike RMS [Stallman], Maddog and the FSF) for only popularizing Linux on the basis of "down with Microsoft." A colleague attended a talk by John "Maddog" Hall, president of Linux International and found the man to spend the whole time attacking Microsoft and its policies.

Big. Fucking. Deal.

Linux must be advertised and promoted on the basis of its abilities, the unique benefits that it presents to businesses and home users alike. That''s the path IBM is taking, and that''s the path I would like to see all Linux/Open Source/Free Software advocates adopt.

quote:
To conclude, Linux is still too young to be used by companies into their information system. Still managers are observing the market, and the fact that IBM communicates on the benefits of Linux on their mainframes (basket ball commercial) does interest a lot of people.

Exactly. Linux will eventually penetrate the marketplace, but only as it matures into a system that presents real, tangible benefits to its users.

[ GDNet Start Here | GDNet Search Tool | GDNet FAQ ]
[ MS RTFM [MSDN] | SGI STL Docs | Boost ]
[ Google! | Asking Smart Questions | Jargon File ]
Thanks to Kylotan for the idea!

Oluseyi:
There is just one trouble with IBM: they practice really high prices for their service (alas I experienced recently that their service is not always of prime quality). Managers know about IBM pricing policy. IMO, I think that even if they are watchful about the latest IBM communication on Linux, they are still dubious on the fact that Linux will become the next generation OS. IBM propose Linux as an OS for two reasons:
- they do not have an OS for small stations.
- they want to propose complete functional solutions at a lower cost than what they charge normally to get a better hold on the market of small companies.

As for Unix, there are many companies who market their own OS (and their OS are not strictly compatible; this could explain why Sun uses Linux as a migration platform):
- IBM -> AIX
- SUN -> SOLARIS
- HP -> HPEX
When you buy an IBM Mainframe, the OS installed is AIX. Do not forget that when you need to create a backup of a server, there are many solutions to mirror data. Still what is recommended is to have separated servers, each server in two different buildings with their own disks. For that kind of backup the main software used and developped is to my knowledge HACMP from IBM (and HAGEO for distances between servers of more than 10 km). It has not been ported to Linux. Why use Caldera''s OS ?

Finally, many companies today use the following combo for computer equipment: Mainframe (IMB, SUN, HP with their proprietary OS) + work stations (often Dell for its lower cost thus Microsoft OS, thus the communicating standard is MS in the world (excel, word, explorer, visual C++, access); this is why an anti microsoft communication will not work: you cannot break a worldwide standard). In medium to large companies, the cost of the work stations is often higher than the cost of the mainframe.

To conclude, for Linux to break through, Linux advocates must convince small companies. I agree with you that an anti microsoft communication will not serve the Linux promotion, it will just give time to competitors to grab that market.

Ghostly yours,
Red.

quote: Original post by Red Ghost
work stations (often Dell for its lower cost thus Microsoft OS, thus the communicating standard is MS in the world (excel, word, explorer, visual C++, access); this is why an anti microsoft communication will not work: you cannot break a worldwide standard).

If you look at how linux is making its way in its almost doing it via the backdoor, with systems like Samba in place to take over file sharing, programs which are compatible with M$ office apps, explorer like interfaces in KDE and Gnome.
However, one point, I am as yet to find any IDE which comes close to what Visual Studio has with regards to layout etc, and with VS.Net M$ have come up with another damn good IDE.

Still, this has all drifted off topic from ''Linux for games'' to ''Linux for the work place''

and to counter one of the things in this about PCs not coming with Linux installed HP recently started shipping PCs with Red Hat installed (Linux Format 26, Page 7) so Joe Six-pack doesnt have a reason to complain any more

quote: Original post by _the_phantom_
If you look at how linux is making its way in its almost doing it via the backdoor, with systems like Samba in place to take over file sharing, programs which are compatible with M$ office apps, explorer like interfaces in KDE and Gnome.

Why would a company or person choose an alternative that isn''t supported and may not be perfectly compatible (especially as new versions of the production software are released) instead of upgrading to the "latest and greatest"? Furthermore, applications like MSOffice and IE are irrelevant in big industry. Only recently have some industrial sector software vendors begun to explore the possibility of porting their products to Linux. Until that happens, Linux is not viable for a large section of big industry - sectors that depend on UNIX or Windows NT.

quote:
and to counter one of the things in this about PCs not coming with Linux installed HP recently started shipping PCs with Red Hat installed (Linux Format 26, Page 7) so Joe Six-pack doesnt have a reason to complain any more

Dell was shipping workstations and desktop PCs with RedHat Linux last year. They quit due to extremely low demand.

[ GDNet Start Here | GDNet Search Tool | GDNet FAQ ]
[ MS RTFM [MSDN] | SGI STL Docs | Boost ]
[ Google! | Asking Smart Questions | Jargon File ]
Thanks to Kylotan for the idea!

quote: Original post by Oluseyi
and to counter one of the things in this about PCs not coming with Linux installed HP recently started shipping PCs with Red Hat installed (Linux Format 26, Page 7) so Joe Six-pack doesnt have a reason to complain any more

Dell was shipping workstations and desktop PCs with RedHat Linux last year. They quit due to extremely low demand.


Well, I guess they have started again then



[edited by - _the_phantom_ on May 8, 2002 3:39:36 PM]

Maybe people missed my other comment, but what about the factor of WineX in preventing developers from creating native Linux games? Is this a good or bad strategy?

I''d personally like to see native games made using native Linux API''s and middleware rather than going through an emulation layer. Why? Because it sort of defeats the purpose of necessitating Linux or improving its popularity.

Several people here have made the point that Linux needs to do things better or more cheaply than Windows, or provide functionality that M$ can not provide. I think the problem with WineX is that it it doesn''t give a compelling reason for people to try something new. Why invest the time to learn a new OS when you can just play a M$ game? In other words, it offers nothing new to the table. Yeah it''s great for people that already run Linux, but in attracting new users I think it is harmful.

I feel I have to jump in here for a comment... I keep reading other posts saying that support is a major issue with Linux and that is why it is not always a viable business os. I must disagree here. Not on the Linux having great support side, it doesn''t. But on the opposite, that you do not get *great* os support from MS for a low cost. I worked on a helpdesk with two of the largest corporations in the world (contract work, gotta love it). We (the company I worked for) had a contract with MS for support for quite a while. When we had an issue we could not solve, we were able to call them. Unfortunately, this was at an outrageous price. And quite often, the solution that came back from the expensive MS techs was, reload the OS. We used that option for about 2 years, then we dropped it. We learned more about Windows/Office/everything. We supported it ourselves. Going to MS is not always a viable option. Going to a contracted helpdesk that knows what they are doing, is a viable option (usually). What I''m saying is, there are helpdesks available that will do Linux support. Going to MS is expensive, for almost all companies. So the issue of support really isn''t an issue in my mind, it''s an oportunity. Someone want to start a linux helpdesk company? We''ll make a lot of money

Always remember, you''''re unique. Just like everyone else.