2011-11-12 Almost every Unix user knows the program top(1). Most of them uses it in one way: fire it up; stare at the reordering of the lines for some time; quit it. Some users might even know how to change the sort column. These days, I learned much more about top. To be exact, I learned about the top from the procps tools (that's the one usually available on GNU/Linux systems). Other top implementaions might not offer much more than the simple usage scheme, described above. Top is a front-end to ps, kill, uptime, etc, for human users only. Hence, you don't need to care for compatibility much. You can use what's available on the system you're on. If you want to use top for system live monitoring, the simple usage pattern from above won't give you enough. But the top from procps offers much more. You just need to explore it. Do so! Read the man page! Try `A', `a'/`w', `z', `x', `<'/`>', `W', `- ', `_' ... It's too much to list here. Read the man page. Then you might find in GNU/Linux's default top implementation most of the stuff you used to install htop for. My vote for the forgotten tool of the month goes to look(1). You call look with a word prefix as argument, and it will return you any words with the same prefix from /usr/share/dict/words. (Actu- ally, it's mainly a wrapper around fgrep, anchored to the begin- ning of the line.) In many cases such a lookup provides enough spelling information and you won't need to query Internet resources. Having the tool look(1) is only half the part; you also need a word list. I took a look, which packages provide /usr/share/dict/words. One is dictionaries-common. Fine, that sounded good. But well, it does not provide a word list, it only provides a symlink, named /usr/share/dict/words. Ahh, here we enter the Debian symlink hell, aka. alternatives, again. Several packages provide actual word lists. For American English spel- ling, you would use `wamerican' or `wamerican-{large,huge,...}'. The alternatives system surely has advantages, and needing sym- link masses for it may be okay, but I simply don't like it. However, look(1) is a fine tool and a nice discovery. http://marmaro.de/lue/ markus schnalke