Tony Asaro says that CAS serve, essentially, two purposes:

  • To serve as an immutable proof for compliance purposes by its inherent nature of WORM property.
  • To convert a data store into an information store by incorporating federated search and indexing.

Will data be seen as information by indexing/searching? Yes!! It sounds cool and new to me. What other inherent properties of CAS are interesting and finds application?

Flexibility vs Capability

John McArthur writes in his blog that flexibility is as valuable as capability.

He suggests to the inventors and innovators:

For entrepreneurs and inventors who are looking for opportunity, I highly recommend that you look for opportunities to unlock chains of inflexibility for companies that are trapped by current applications, systems or processes.

And to the person in charge of supplier selection process:

For buyers of technology, before adopting any new technology, always remember to consider these two flexibility factors:

  1. What is the financial and operational impact, if the supplier or product ceases to exist?
  2. If my current supplier with whom you have entrusted your information, your applications, or your processes, ceases to perform to your satisfaction, how do you migrate back to your own environment or to another supplier?

There are two things to learn from this piece.

Perl facts

Some facts about perl that I learnt today:

  • References are scalars. It is the third scalar, the first two being strings and numbers.
  • Also, references are of two types. Symbolic references just refer to a variable name and not the contents per se. Hard references refer directly to the value. One can compare this to the symbolic and hard references we find in file system.
  • One thing to note in perl is that non-scalar data structures like arrays and hashes are linear. That is they can hold only scalars and not arrays and hashes. So when we refer to array of arrays, array of hashes etc., we merely mean that the array contains references to arrays or hashes, but not arrays or hashes themselves.
  • References can be created for anonymous scalars, or constants like 7, “nagp” also.
  • A reference can be created by merely pre-pending a \ to the variable or to the anonymous version.
  • References to anonymous arrays are created with [ and ], instead of with normal ( and ).
  • References are dereferenced with a ->.
  • References to anonymous hashes are created with { and }. with key => value pairs separated by commas. The value can be another mess of data structures.
  • Perl does no implicit referencing or dereferencing. When a scalar holds a reference, it just behaves like a scalar and does not magically do any auto mapping.
  • If we know what the dereferenced type is, then we can dereference the scalar. For example, if we know $foo is a reference to an array, then that array can be referenced as @$foo. If it is a hash, it can be as %$foo. And if it is just a scalar, then it is $$foo.

/dev/random and /dev/urandom

If you really don’t care about so much entropy in the random numbers generated, it is recommended to use /dev/urandom instead of /dev/random. Because, reading from /dev/random hangs forever if it is not able to accumulate enough entropy.

Lost root password?

Sometimes it so happens that we forget root password. And even more painful thing is that when we boot in a single user mode, with a hope that it will give us a root prompt, it also asks for a password.

To get rid of this problem, do the following:

  • When the GRUB prompt shows up, instead of letting it boot, select the kernel and edit it.
  • At the kernel command line, instead of giving single, give init=/bin/sh. This makes the kernel to start sh instead of the actual init process. Note that parameters like single etc. should not be given because sh does not understand them resulting in a panic.
  • When the root prompt shows up, give a passwd command and change the password.