This story reminds me of another paradox: the anonymous paradox, where people feel like it is more "anonymous" to order online stuff that they don't want people to know about. But actually, if you really want this hardcore XXX movie but you don't want people to know about it, you should go physically to the adult store and pay cash instead of leaving an electronic paper trail. (Same rule applies to the purchase of Celine Dion's latest album!).
Instead of a NAS, I use two Antec 900 cases with low-end pc. Each case can hold up to 9 HD (if you don't need an internal DVD), and the disks are located in 3-disk containers with a dedicated 120mm fan (yep, one fan for every 3 disks!). There is also a huge 200mm fan on top of the case and a 120mm on back. With all those fans the disks stay cool no matter how badly you ride them, and the fans can be set at the minimum speed; there is not much noise. Also there is 2xUSB and 1xfirewire ports on top of the case, which I use for the O/S.
So in a single case (which is also quite the looker) you can get 9x500GB or even 9x1TB. Of course you need to find a mobo with enough SATA (or IDE if you prefer) connectors, but 2x SATA RAID 1 cards are cheap and reliable. And you also need a good PSU (I live and die by Antec!).
I don't know where you live, but here in Canada this whole setup is quite cheap: -mobo+cpu+2GB DDR2: 225$ -psu: 100$ -SATA RAID cards (2): 50$ -Antec 900: 125$ -9x500GB HDs: 800$ -USB stick (for the O/S): 20$
So for less than 1500$ you get a 2.2TB fully redundant storage, on which you can connect using Samba, NFS or whatever protocol your Linux O/S supports. As for myself, I use iSCSI and LVM in my client PC to connect to my 2 Antec servers so my system is completely redundant.
The only tricky part is to access the RAID cards from Linux, but even with no-name brands you can make it work with stock drivers and a good search engine...
This is a very good advice. Being a One-Man-Shop is excellent in the early years of a IT career, this allows you to see what are your actual interests and natural skills. But over the years, a better focus will bring a better pay and more career opportunities, because the resume becomes more attractive to good employers.
This does not mean that continuous improvement or learning is bad. It's great to keep an open mind, but a focus on a specific set of skills does provide great benefits.
As for being bored... I found out that most of the time a little initiative will bring more interest to a job. If you know your job well, you can find some motivation in improving your own productivity, which might include making suggestions to slightly improve things that are not entirely under your own control (and not getting pissed not to see your suggestions implemented). This shows that you care about your job and your company.
In the early years of Standard Oil, employees were invited to find ways to improve their own processes, up to the point where they could get someone with less skills to do the job; then they would move to higher positions. This is the best career advice I ever found.
You obviously don't know what the CISSP is. It is not a technical certification, it is aimed at people who are interested in risk assessment, security management, and such. So people doing a CISSP are very likely to already be part of the management.
The certification process is far more complicated than you seem to think. First of all, just to qualify for the exam you need to have a proven 4-year experience in IT security (or 3-year + college degree). Then if you pass the exam, which is not that easy, you still have to go through an interview where your professional experience is validated by people who are not that easy to bullshit. You also need to provide some kind of credentials or endorsement. It's a serious business.
So knowing the difference between a rootkit and spyware might be interesting in a very small segment of the certification process. But the actual requirement is to master the ISO-19977 guidelines, and to have a proven track record in information security management.
I use prepaid minutes for my cellphone (cheap and anonymous). When I want to refill my account I must purchase a card, then call my provider and give the card PIN to add minutes to my account. It used to be a quick and efficient punch-in system, but they switched to voice recognition a year ago and it is now very painful.
They greatly improved the thing, like allowing to punch-in the password and the PIN (which were almost never recognized correctly by the system). They also removed options that are too similar, like "Prepaid card" or "Credit card". But they keep changing the navigation so you cannot interrupt the messages otherwise you get lost in the menus, so it takes like 2 or 3 minutes instead of 20 seconds.
In any case, it has been a year and I still feel like an idiot when I talk to the system so I try not to do it in public. My only pleasure is to give a rude and obscene answer to the last question of the system, which is "Can I do something else for you" (although I hang up before being transfered to a TSR)...
It might violates the laws of thermodynamics according to our current scientific achievements. But science has a way of proving itself wrong over time... so I'd rather believe in magic than state that if there was such a thing as aliens then their spaceship would definitely have "heat vent". But that's just me!
-Visual Basic is not a scripting language; the programs are compiled ("true" EXE since version 5). Two stripped-down version of Visual Basic are available though; VBA (Visual Basic Application), a scripting language designed mostly for Microsoft Office, and VBScript, which is an actual scripting language (for ASP or Windows Scripting Host mostly).
-Most of the Office (and some of the Windows) libraries are still unavailable in.NET; if you use the framework to interact with them you have to use interop, which means that you basically cast everything and make no use of.NET features.
I completely agree. A lot of clerks or people working in warehouse are used to punch in numbers, press RETURN and move to the next item. Switching to a mouse-based system would not be efficient, people would complain all the time and ask to "go back like it was".
In real life you have to adapt the machine to the people who are using it, not the opposite. Unless you really want SAP...
The laptop goes everywhere [...], the PDA does, too
Why do you need a PDA if you have your laptop with you all the time? If you really, really want a small and clunky GUI for your address book or calendar, I'm sure you can find a PDA emulator somewhere.
From your post I understand that these people are using Exchange 5.5, which is the last relase not integrated in Active Directory.
Here are my suggestions:
1) Since they must already have an Outlook license, they could use Exchange 2003 in RPC/SSL. This is 100% Active Directory integrated, and they don't need a web client since Outlook can connect directly. Exchange itself is not so expensive, however they must also upgrade the CALs, which can boost the bill for 6k users.
2) A more reliable solution: Postfix and imapd. I strongly suggest Cyrus Imap as it saves each email as a separate file, which is incredibly efficient for backups and archives. For the frontend, most webmail are fine (Horde, SquirrelMail, Neomail) and includes the calendar stuff; https can be setup easily. This kind of setup can use Active Directory for the users authentication and can run on multiple servers (clustered or not). Note: in this setup the users could still use Outlook for which they paid a shitload of money.
A few notes: -Users will complain if they have to move from Outlook to whatever web interface.
-Antispam and antivirus software are much cheaper and much more efficient if they go with the Postfix scenario.
-6k remote users is a lot, especially if they use this email as a primary business address. Therefore the IT dept must act as a small ISP. Why not calling a few ISP in the area and see what is their setup?
-If they go with the Postfix scenario, if they use x86 hardware I would suggest Linux with a recent kernel. However, if they have the opportunity to purchase hardware, I would strongly suggest Solaris 10 and a sparc machine (possibly two for a better uptime). At the present moment, I can't think of a better os/hardware match for a mail server.
Recently I've been installing Microsoft CRM 3.0 for a customer, and I found the implementation guide quite puzzling.
Actual requirement: "You must be logged in with Domain Administrator and Local Administrator privileges when running Microsoft CRM Setup".
Also, two Active Directory requirements that look suspiciously mutually exclusive: -Active Directory must be in native mode before you can install Microsoft CRM. -For the Microsoft CRM servers to have access to Active Directory Organizational Units where users are located, you will add the following accounts to the Pre-Windows 2000 Compatible group in Active Directory: (various groups listed)
Other odd points: -The computer name of the Microsoft SQL Server cannot contain an underscore (_) -Microsoft CRM Server is supported only with a default instance of SQL Server. Named instances of SQL Server are not supported with Microsoft CRM. -Microsoft CRM cannot access SQL Server with TCP/IP, only named pipes. -The setup won't succeed if the "Error reporting" service is stopped. That is the service that is giving the "opportunity" to send error messages to Microsoft when something goes wrong.
This setup was looking very suspicious to me, so I prepared a risk assessment and required the customer to sign it before I installed the package.
Many country have much more severe regulations against encryption technologies than the USA. Just a few years ago there was an incredibly severe law in France, stating that even domestic encryption was under the government control.
Believe it or not, you actually live in a (somehow) free country.
If you suspect that your company is doing something against the law, it is your duty to raise the issue with your boss. You can refer him to the section 12.1 ("Comply with legal requirements") of the ISO recommended information security practices (ISO/IEC 17799). But be sure that there is wrongdoing; as an example, in Canada you are free to put secret cameras in the workplace as long as you can demonstrate in court that you are protecting equipments and not watching specifically someone. It is much trickier if there is audio involved.
Once you have raised the issue, if your boss explains that the cameras are there for a general security purpose, then it is up to you to evaluate if you are comfortable with the situation; you may stay and be annoyed, or leave. But if your boss does not acknowledge the issue, or if his behavior is not acceptable, walk away from this place.
In both case you should not involve the authorities, unless you feel that there is something very, very wrong. Revenge and bitterness are pointless and might just come in your way someday.
As for the cameras themselves, I think it is a big investment for what it is worth. The company must deal with storage and handling of the media (tape or DVD); if any incident occur they must go back in time, search for a while, then hope the picture is clear, and in most case it will have no value in court. It's a lot of annoyance for little ROI. But anyway I don't mind being on camera myself, as long as pictures of me spilling coffee in the fax server don't get on "America's Funniest Videos".
MS-Project is the way to go. Here are a few pointers:
-At first, don't bother with dates. Use only the priority, % completed and estimated duration (always in hours). Don't be shy to remove columns and the gantt chart from the task entry form; at first it is more of an annoyance, especially if you manage many small projects in the same file. -Add all working people in the resources sheet and set an average cost by hour. If you want to manage the schedule (and not only the overall cost), don't forget to give an accurate availability of the resources (what % of their time will be committed to the project). As an example, if your sysadmin is part of a project, don't commit him for more than 25% of its time. -Add all purchase (actual or planned) in the resources sheet, and affect them to the tasks when needed. Don't forget to set the price. -Add the WBS column in the task entry form, and use plenty of subtasking (increasing or decreasing indent will create child tasks). This will help you to group tasks (and cost). -While the various projects are moving, use the filters in the Project menu to hide the completed tasks, or the ones with a priority far from what you can afford on day to day. This will help you to keep track of what is really important. Reapply the filters once in a while, as they don't stick with the new values. -Use plenty of notes in the tasks. These can help you remember why a specific project is on hold.
Most important of all: store the project in a database, not in a MPP file. This can be done with the ODBC button in the File-Save dialog. On the first save MS-Project will create all the tables in the selected database. Then you can connect to this database and do all kinds of wonder with the data, like reports, real-time monitor, on-request status report via web services, and so on. In you DB you can add tables for additional information about purchases: supplier, warranty, billing, and so on, which you can fill either from a plugin in MS-Project or directly in the DB. You don't need Project Server to do it.
Actually, unless you are ready to invest in a lot of work, my advice is that you should not try to setup Project Server or Outlook integration. It is probably not worth the effort.
The best setup I found out is to save the MS-Project data in SQL Server, and build web reports with Reporting Services (included with the SQL license). With Reporting Services people can watch live reports, or schedule email-delivery in any format they want. As an example people could query the system for the tasks they are assigned to and get it automatically in a spreadsheet every morning, to help them work out a good daily schedule. Or the management could get an automated status report with charts and everything, without having to ask it from you.
Then you could either update the % completed in MS-Project, or build a small web application to let people do it themselves.
MS-Project can be a very handy tool, much more convenient than Excel for budgeting and planning. Just don't let people or features scare you, take the time to play around and don't hesitate to heavily customize your entry forms. Expand it with additional stuff in the database, build web forms, tools and reports. As long as you keep the data up-to-date, it will work.
Until now I was sure that the only thing with a higher TCO than Oracle was a Sea Stallion helicopter (38 hours of maintenance required for every hour of flight). I guess I never thought about SAP TCO because most of the SAP rollouts I heard about failed.
Those projects are so incredibly expensive, I have no idea what kind of scale they use to calculate the TCO. Teradollars? I can imagine a board meeting (CIO: "Hey guys, we must make room for 317 Teradollars in the next budget for this SAP thingy. So I guess we'll have to forget about the Winzip licenses for now.").
Seriously, a friend of mine is convinced that SAP is part of a secret plan to crush the western economy.
You mix up "duty" and "role". While the role of IT department is to provide IT services to the enterprise, its first duty is to ensure security and stability of existing information systems. Just like the doctor's first duty is "first, do no harm".
Accounting is also a service industry. However, accountants won't dip in the cashflow every time a user heard about a great investment strategy from some guy on CNN. Why? Because their duty is to safeguard the financial well-being of the company, and unless told otherwise by the top management they will keep the cashflow just the way they think it must be.
This being said, IT is not a service to people. It is a service to the company.
The core problem is Corporate IT departments not spending the time to explore possibilities in new products - their knee-jerk reaction is 'no' (for many reasons).
The first duty of "Corporate IT departments" is to ensure security and stability of systems. Therefore if a product can potentially become a security breach, it has to be banned.
Then if users really want the banned software, THEY must build a business case and convince management that the benefits are sufficient to invest IT time in a formal risk-assessment process. Then, once they have the pros and cons, the people in management can make a decision.
Slashdot Mirror
This story reminds me of another paradox: the anonymous paradox, where people feel like it is more "anonymous" to order online stuff that they don't want people to know about. But actually, if you really want this hardcore XXX movie but you don't want people to know about it, you should go physically to the adult store and pay cash instead of leaving an electronic paper trail. (Same rule applies to the purchase of Celine Dion's latest album!).
Instead of a NAS, I use two Antec 900 cases with low-end pc. Each case can hold up to 9 HD (if you don't need an internal DVD), and the disks are located in 3-disk containers with a dedicated 120mm fan (yep, one fan for every 3 disks!). There is also a huge 200mm fan on top of the case and a 120mm on back. With all those fans the disks stay cool no matter how badly you ride them, and the fans can be set at the minimum speed; there is not much noise. Also there is 2xUSB and 1xfirewire ports on top of the case, which I use for the O/S.
So in a single case (which is also quite the looker) you can get 9x500GB or even 9x1TB. Of course you need to find a mobo with enough SATA (or IDE if you prefer) connectors, but 2x SATA RAID 1 cards are cheap and reliable. And you also need a good PSU (I live and die by Antec!).
I don't know where you live, but here in Canada this whole setup is quite cheap:
-mobo+cpu+2GB DDR2: 225$
-psu: 100$
-SATA RAID cards (2): 50$
-Antec 900: 125$
-9x500GB HDs: 800$
-USB stick (for the O/S): 20$
So for less than 1500$ you get a 2.2TB fully redundant storage, on which you can connect using Samba, NFS or whatever protocol your Linux O/S supports. As for myself, I use iSCSI and LVM in my client PC to connect to my 2 Antec servers so my system is completely redundant.
The only tricky part is to access the RAID cards from Linux, but even with no-name brands you can make it work with stock drivers and a good search engine...
This is a very good advice. Being a One-Man-Shop is excellent in the early years of a IT career, this allows you to see what are your actual interests and natural skills. But over the years, a better focus will bring a better pay and more career opportunities, because the resume becomes more attractive to good employers.
This does not mean that continuous improvement or learning is bad. It's great to keep an open mind, but a focus on a specific set of skills does provide great benefits.
As for being bored... I found out that most of the time a little initiative will bring more interest to a job. If you know your job well, you can find some motivation in improving your own productivity, which might include making suggestions to slightly improve things that are not entirely under your own control (and not getting pissed not to see your suggestions implemented). This shows that you care about your job and your company.
In the early years of Standard Oil, employees were invited to find ways to improve their own processes, up to the point where they could get someone with less skills to do the job; then they would move to higher positions. This is the best career advice I ever found.
Don't choose the company that has no respect for the privacy of the people using its free email service.
anyway the money will come from the same place :-)
You obviously don't know what the CISSP is. It is not a technical certification, it is aimed at people who are interested in risk assessment, security management, and such. So people doing a CISSP are very likely to already be part of the management.
The certification process is far more complicated than you seem to think. First of all, just to qualify for the exam you need to have a proven 4-year experience in IT security (or 3-year + college degree). Then if you pass the exam, which is not that easy, you still have to go through an interview where your professional experience is validated by people who are not that easy to bullshit. You also need to provide some kind of credentials or endorsement. It's a serious business.
So knowing the difference between a rootkit and spyware might be interesting in a very small segment of the certification process. But the actual requirement is to master the ISO-19977 guidelines, and to have a proven track record in information security management.
I use prepaid minutes for my cellphone (cheap and anonymous). When I want to refill my account I must purchase a card, then call my provider and give the card PIN to add minutes to my account. It used to be a quick and efficient punch-in system, but they switched to voice recognition a year ago and it is now very painful.
They greatly improved the thing, like allowing to punch-in the password and the PIN (which were almost never recognized correctly by the system). They also removed options that are too similar, like "Prepaid card" or "Credit card". But they keep changing the navigation so you cannot interrupt the messages otherwise you get lost in the menus, so it takes like 2 or 3 minutes instead of 20 seconds.
In any case, it has been a year and I still feel like an idiot when I talk to the system so I try not to do it in public. My only pleasure is to give a rude and obscene answer to the last question of the system, which is "Can I do something else for you" (although I hang up before being transfered to a TSR)...
It might violates the laws of thermodynamics according to our current scientific achievements. But science has a way of proving itself wrong over time... so I'd rather believe in magic than state that if there was such a thing as aliens then their spaceship would definitely have "heat vent". But that's just me!
You got things wrong.
.NET; if you use the framework to interact with them you have to use interop, which means that you basically cast everything and make no use of .NET features.
-Visual Basic is not a scripting language; the programs are compiled ("true" EXE since version 5). Two stripped-down version of Visual Basic are available though; VBA (Visual Basic Application), a scripting language designed mostly for Microsoft Office, and VBScript, which is an actual scripting language (for ASP or Windows Scripting Host mostly).
-Most of the Office (and some of the Windows) libraries are still unavailable in
The Microsoft Way:
1) For small problems: reboot
2) For medimum problems: reinstall
3) For big problems: upgrade
RTFA.
"Before the regular cries of 'Supply Ubuntu' get too loud - that will _not_ work."
I completely agree. A lot of clerks or people working in warehouse are used to punch in numbers, press RETURN and move to the next item. Switching to a mouse-based system would not be efficient, people would complain all the time and ask to "go back like it was".
In real life you have to adapt the machine to the people who are using it, not the opposite. Unless you really want SAP...
Why do you need a PDA if you have your laptop with you all the time? If you really, really want a small and clunky GUI for your address book or calendar, I'm sure you can find a PDA emulator somewhere.
The harder you work, the lesser you get paid.
From your post I understand that these people are using Exchange 5.5, which is the last relase not integrated in Active Directory.
Here are my suggestions:
1) Since they must already have an Outlook license, they could use Exchange 2003 in RPC/SSL. This is 100% Active Directory integrated, and they don't need a web client since Outlook can connect directly. Exchange itself is not so expensive, however they must also upgrade the CALs, which can boost the bill for 6k users.
2) A more reliable solution: Postfix and imapd. I strongly suggest Cyrus Imap as it saves each email as a separate file, which is incredibly efficient for backups and archives. For the frontend, most webmail are fine (Horde, SquirrelMail, Neomail) and includes the calendar stuff; https can be setup easily. This kind of setup can use Active Directory for the users authentication and can run on multiple servers (clustered or not). Note: in this setup the users could still use Outlook for which they paid a shitload of money.
A few notes:
-Users will complain if they have to move from Outlook to whatever web interface.
-Antispam and antivirus software are much cheaper and much more efficient if they go with the Postfix scenario.
-6k remote users is a lot, especially if they use this email as a primary business address. Therefore the IT dept must act as a small ISP. Why not calling a few ISP in the area and see what is their setup?
-If they go with the Postfix scenario, if they use x86 hardware I would suggest Linux with a recent kernel. However, if they have the opportunity to purchase hardware, I would strongly suggest Solaris 10 and a sparc machine (possibly two for a better uptime). At the present moment, I can't think of a better os/hardware match for a mail server.
After scanning your comment, I had to change a fuse on my bullshit detector.
Last time this kind of thing happened, my detector was left open for an entire episode of O'Reilly on Fox News.
Recently I've been installing Microsoft CRM 3.0 for a customer, and I found the implementation guide quite puzzling.
Actual requirement: "You must be logged in with Domain Administrator and Local Administrator privileges when running Microsoft CRM Setup".
Also, two Active Directory requirements that look suspiciously mutually exclusive:
-Active Directory must be in native mode before you can install Microsoft CRM.
-For the Microsoft CRM servers to have access to Active Directory Organizational Units where users are located, you will add the following accounts to the Pre-Windows 2000 Compatible group in Active Directory: (various groups listed)
Other odd points:
-The computer name of the Microsoft SQL Server cannot contain an underscore (_)
-Microsoft CRM Server is supported only with a default instance of SQL Server. Named instances of SQL Server are not supported with Microsoft CRM.
-Microsoft CRM cannot access SQL Server with TCP/IP, only named pipes.
-The setup won't succeed if the "Error reporting" service is stopped. That is the service that is giving the "opportunity" to send error messages to Microsoft when something goes wrong.
This setup was looking very suspicious to me, so I prepared a risk assessment and required the customer to sign it before I installed the package.
Here is a nice web-based solution: http://iptrack.sourceforge.net/
...
We are using it at the office and it is very handy.
There is a lot of features, including DNS management, search tools, routing tables management,
No shame here, it's a great project! Keep up the good work.
Many country have much more severe regulations against encryption technologies than the USA. Just a few years ago there was an incredibly severe law in France, stating that even domestic encryption was under the government control.
Believe it or not, you actually live in a (somehow) free country.
If you suspect that your company is doing something against the law, it is your duty to raise the issue with your boss. You can refer him to the section 12.1 ("Comply with legal requirements") of the ISO recommended information security practices (ISO/IEC 17799). But be sure that there is wrongdoing; as an example, in Canada you are free to put secret cameras in the workplace as long as you can demonstrate in court that you are protecting equipments and not watching specifically someone. It is much trickier if there is audio involved.
Once you have raised the issue, if your boss explains that the cameras are there for a general security purpose, then it is up to you to evaluate if you are comfortable with the situation; you may stay and be annoyed, or leave. But if your boss does not acknowledge the issue, or if his behavior is not acceptable, walk away from this place.
In both case you should not involve the authorities, unless you feel that there is something very, very wrong. Revenge and bitterness are pointless and might just come in your way someday.
As for the cameras themselves, I think it is a big investment for what it is worth. The company must deal with storage and handling of the media (tape or DVD); if any incident occur they must go back in time, search for a while, then hope the picture is clear, and in most case it will have no value in court. It's a lot of annoyance for little ROI. But anyway I don't mind being on camera myself, as long as pictures of me spilling coffee in the fax server don't get on "America's Funniest Videos".
MS-Project is the way to go. Here are a few pointers:
-At first, don't bother with dates. Use only the priority, % completed and estimated duration (always in hours). Don't be shy to remove columns and the gantt chart from the task entry form; at first it is more of an annoyance, especially if you manage many small projects in the same file.
-Add all working people in the resources sheet and set an average cost by hour. If you want to manage the schedule (and not only the overall cost), don't forget to give an accurate availability of the resources (what % of their time will be committed to the project). As an example, if your sysadmin is part of a project, don't commit him for more than 25% of its time.
-Add all purchase (actual or planned) in the resources sheet, and affect them to the tasks when needed. Don't forget to set the price.
-Add the WBS column in the task entry form, and use plenty of subtasking (increasing or decreasing indent will create child tasks). This will help you to group tasks (and cost).
-While the various projects are moving, use the filters in the Project menu to hide the completed tasks, or the ones with a priority far from what you can afford on day to day. This will help you to keep track of what is really important. Reapply the filters once in a while, as they don't stick with the new values.
-Use plenty of notes in the tasks. These can help you remember why a specific project is on hold.
Most important of all: store the project in a database, not in a MPP file. This can be done with the ODBC button in the File-Save dialog. On the first save MS-Project will create all the tables in the selected database. Then you can connect to this database and do all kinds of wonder with the data, like reports, real-time monitor, on-request status report via web services, and so on. In you DB you can add tables for additional information about purchases: supplier, warranty, billing, and so on, which you can fill either from a plugin in MS-Project or directly in the DB. You don't need Project Server to do it.
Actually, unless you are ready to invest in a lot of work, my advice is that you should not try to setup Project Server or Outlook integration. It is probably not worth the effort.
The best setup I found out is to save the MS-Project data in SQL Server, and build web reports with Reporting Services (included with the SQL license). With Reporting Services people can watch live reports, or schedule email-delivery in any format they want. As an example people could query the system for the tasks they are assigned to and get it automatically in a spreadsheet every morning, to help them work out a good daily schedule. Or the management could get an automated status report with charts and everything, without having to ask it from you.
Then you could either update the % completed in MS-Project, or build a small web application to let people do it themselves.
MS-Project can be a very handy tool, much more convenient than Excel for budgeting and planning. Just don't let people or features scare you, take the time to play around and don't hesitate to heavily customize your entry forms. Expand it with additional stuff in the database, build web forms, tools and reports. As long as you keep the data up-to-date, it will work.
Until now I was sure that the only thing with a higher TCO than Oracle was a Sea Stallion helicopter (38 hours of maintenance required for every hour of flight). I guess I never thought about SAP TCO because most of the SAP rollouts I heard about failed.
Those projects are so incredibly expensive, I have no idea what kind of scale they use to calculate the TCO. Teradollars? I can imagine a board meeting (CIO: "Hey guys, we must make room for 317 Teradollars in the next budget for this SAP thingy. So I guess we'll have to forget about the Winzip licenses for now.").
Seriously, a friend of mine is convinced that SAP is part of a secret plan to crush the western economy.
You mix up "duty" and "role". While the role of IT department is to provide IT services to the enterprise, its first duty is to ensure security and stability of existing information systems. Just like the doctor's first duty is "first, do no harm".
Accounting is also a service industry. However, accountants won't dip in the cashflow every time a user heard about a great investment strategy from some guy on CNN. Why? Because their duty is to safeguard the financial well-being of the company, and unless told otherwise by the top management they will keep the cashflow just the way they think it must be.
This being said, IT is not a service to people. It is a service to the company.
The first duty of "Corporate IT departments" is to ensure security and stability of systems. Therefore if a product can potentially become a security breach, it has to be banned.
Then if users really want the banned software, THEY must build a business case and convince management that the benefits are sufficient to invest IT time in a formal risk-assessment process. Then, once they have the pros and cons, the people in management can make a decision.