A little exercise using a sequence of bash shell commands (I'm including the "$" shell prompt, to make it clear which lines are actual commands; the lines without "$" are either keyboard input or command output):

$ cat <<EOF > test.in
one o n e
two t w o
three t h r e e
EOF

$ perl -pe 's/(\w+)/\U$1/' < test.in > test.out

$ cat test.out
ONE o n e
TWO t w o
THREE t h r e e
[download]

If you want to store the one-line perl script as an executable file (so that you don't have to type  perl -pe 's/(\w+)/\U$1/' every time you want to up-case the first word of every line in a data stream), you would do this:

$ cat <<EOF > upcase-firstword
#!/usr/bin/perl -p
s/(\w+)/\U$1/;
EOF

$ chmod +x upcase-firstword

$ upcase-firstword < test.in > test2.out

$ diff test.out test2.out
[download]

Move the "upcase-firstword" file to some directory in your PATH, and you can run it no matter which directory you are in at the moment. (You can give it a shorter name if you like.)

Using the "cat" command to write a perl script like that can be lots of fun, although many programmers would not consider it to be their preferred method for programming.