Following the Joel Test for SysAdmin; For all the places don't meet the test, what needs to be done, and in what order?

Following The Joel Test for SysAdmin there have been some great suggestions;

but imagine for a sec that you're already working for someone who doesn't meet the test, or you're being asked to bring an organisation up to this standard

1) What are the most important issues to address?
2) In what order would you address them?

From serverfault Pixelicious

• I start with asking for documentation about the current business processes and workflows. If that is not available I request it to be mapped, if that is not possible I create these documents albeit only from the IT perspective.

  I will let that be reviewed and based on that I create the SLA. Using these documents I can choose the right tools, this will result in a dependency diagram of software, hardware and support. Based on that I can create the budget.

  If the budget is not maintainable, the SLA has to be adapted to reflect that.

  From Martin P. Hellwig

• Documentation and backup and restore, for sure. If you don't have those you're screwed.

  Clear out any ageing hardware, ageing software, or generally cranky old boxes, if possible. Those 20 hours a week keeping "old Betsy" up and running could be more productively spent elsewhere.

  A top to bottom review of practices and procedures. If something is being done wrong, "but we've always done it that way" is not good enough.

  Keep it in your mind that your objective should be to make things as boring as possible. You want to be able to come to work in the morning and say "oh YAWN, everything's still up and running smoothly". How does each item you're going to change move you closer to this goal?

  Pixelicious : The idea of making everything as boring as possible is a funny, but insightful one :) I like it!
  From mh