First of all, a quick introduction training: I started tinkering with Linux and free software in 1998, and began to use them professionally in 2001. It was a season complicated by a lack of information and documentation available, besides the absence of free solutions that make life easier for everyone and are common today.
The mere idea of installing Linux was already an exciting interface: text-mode installation, you had to select all the packages manually, and invariably you'd have to download a video module or sound and recompile the kernel if you wanted to everything worked perfectly.
Since then much has changed, but three factors still seem far from change: the distrust of managers to adopt free solutions, the barrier created by the users during the migration, and the flurry of technicians, who believe that a system migration within a company can be just as at home, no studies or tests.
Migrations and deployments within an enterprise is a complex issue, and not because you could install Linux on your home that you can sell the idea of a complete migration within the company where you work. Remember, not only the reputation of Linux within the company that is at stake, but your job is, too. For this task, it takes much more than technical knowledge: we must understand the company, knowing the bureaucracy, and learn to circumvent.
It seems fun, but still see some of the same doubts and errors that I committed in 2001. There should be none. Linux is much easier to install and use, OpenOffice - BrOffice for Brazilians - is far more compatible with Microsoft Office, and several distributions already have programs that are part of the configuration of servers quickly. But it is still common to hear stories of companies that have had problems trying to migrate to free software, mostly caused by internal bureaucracy or outright flurry of Linux enthusiasts.
Thinking about it, I have compiled some of the tricks I learned during my professional life in order to help the beginner. Most are obvious things, but I reiterate: I still see it happening. Just get on mailing lists or chat communities you will find some of the cases treated here.
Without further ado, let the tips.
Learn to speak the language of the people
There is an arrogance common in computer science, and especially among free software users, to find that people have an obligation to know what you're talking about. Not quite.
Each has their ways of understanding things, and not take it into account at most will leave you frustrated, thinking that their superiors or managers from other sectors do not understand the wonder that is the Free Software. And in companies where there are structures and well-defined positions, a "no" to a boss can kill the project early on. To avoid this, it is important to know who you're dealing with and show them the benefits of migration.
The examples are many. If your boss does not understand computer, what good is talking so enthusiastic about that project proxy with several layers of security, user authentication, anti-virus and firewall and integrated access control? In the end you will ask "what do I have to do with it?". In this case, really, do not explain what you will do, but what the company will earn.
If the boss just looks at the financial side, make a table or a report showing how much you spend on unnecessary shipping or how long a user is stationary when the machine is hacked or caught a virus, and demonstrate that this change will decrease, causing overall economy. If your boss is addicted to control and productivity, show another side: that the proxy will generate reports for navigation and use of the Internet, allowing users to identify bottlenecks or unproductive. Okay, without going into any technical detail, you can convince your boss. After all, he is now able to see the benefits in change.
Never say that Linux is "free"
Seriously. I'm not kidding. I must repeat that: Never say that Linux is "free."
Reason? For many people, even some who work with computers, there is a simple and unchanging in the world: the more expensive the product, he will have more quality. If we stop to think outside the computer world, it makes some sense. More expensive products usually denotes higher quality or more features. Now, as a person with that value judgments could face the news that a product is "free"? Probably thinking (or speaking) "For free? So it must be crap ... "
And there's another problem: the claim that a solution is free, you are giving room for those who cheer against change might stop the project if any expense to appear (different equipment, books, help from others, etc...) Then you will explain that extra network card actually has nothing to do with free software, or the module to connect that system to the system owner free is paid, and so on.
When in doubt, always use "free" or "open".
Before starting the migration, check that it is possible and if you know how to do
Something very common in mailing lists and forums on Linux (actually, any BOM), and even the comments of my blog is to find a message something like this:
"Glra speech, I'm here with a pepinão, got a Linux migration project in a company with 160 computers and five servers that need to authenticate via LDAP, through a VPN to a proxy enclosed inside a firewall and plugged into a radio Cell. I talked to my boss who gave everything to migrate to Linux, and now he's on top of me, charging me, and has said she'll send me away if I did not deliver everything tomorrow. I saw on the net that I could do it, but am having some problems. Somebody help me? I stopped at the part where the installation of Linux ask me to restart the computer .... "
Then the guy obviously can not, but it's the boss will get the impression that everything in Linux is more difficult or impossible.
Of course, each case is different: I've seen asking heads of technical things that they simply did not know how. But if the idea of migration came from you, unless you are sure you know what you're doing.
How? Set up a mini-network somewhere in your office, off-grid "official" with two or three machines linked, and do all possible tests. Not because you got a manual on the internet claiming to be "super simple" to end up being simple. Even more if necessary changes in configuration scripts. If you do not know at least how the basic parameters of Samba should not promise an SMB with LDAP authentication for tomorrow.
Ensure that the migration is transparent to the average user
Imagine the scenario: You spent the night installing OpenOffice on all company computers, and early in the morning he gets a call from an official of the Financial, saying it can not correctly open any spreadsheet in Office 2007. You say you'll have to migrate the worksheets, but it will take a while. Five minutes, your boss calls you, telling you to reinstall Office 2007 on the machine employee, since he can not work. The news spreads, and in less than a week everyone is using MS-Office again, with OpenOffice leaning in some corner of the hard disk.
It'sa common good, believe me. Now imagine an employee arriving at work on Monday and finding the login screen of Ubuntu on his computer.
The only way to avoid this is to ensure that the user knows what is happening and to anticipate his needs. The proxy will be completely closed, only allowing access to some sites, but you know what sites each sector tends to go through? The user knows that he must warn you if you can not access the page, or you'll just find that she is off the air? All doors will be closed by the firewall, but no program or website that connects to a specific port? The most important documents are fully compatible (visual, formulas, macros, etc.).? If not, is there any alternative?
The list goes on. Note that this case is not only the technical knowledge that is necessary, but knowledge of the company and the relationship with employees. Well, nobody said it was easy to migrate ...
Know the time to give the arm twisting
Part of the advantage of getting older is knowing how to choose which battles to fight. And that means knowing when to stop the migration project that is stalled for months on end and think about alternatives.
You can not use Linux as standard in all computers? But the servers? OpenOffice has not been approved by all sectors? And the antivirus? So on.
Continue giving blow a knife will not solve anything. At best, you will be known as the fanatic who wants to make the company lose money with such a Linux ...
