>>Third and most provocative, why isn't mdb as easy to use and generally >>useful as solarcat? > > This isn't a very useful comment without getting into some specifics.
Just my thoughts/opinions as someone that often uses both tools to solve problems: I'd suggest that the two tools are designed by 2 different types of folks for 2 different purposes (not a bad thing necessarily): -solariscat - (at least originally as FM was) designed by phone/field support folks to figure out enough about the random kernel core file that was received (i.e. various panics, hard hang or just hung enough for admin to think it was hung and forced core, system resource issue like mem/cpu/net so low/slow that it seemed hung, cores that were 6 months old the customer never realized was there when the machine paniced and recovered in teh middle of then night, etc) and I recall Paul Faramelli asking people "What do you need to know first and foremost when you get a core file from a customer?" during his early days designing the tool. It's as if it was designed to be user friendly and quickly get you to the point that you knew what was broke or had enough to file a bug. -mdb - designed by engineers that (usually) know the bits-n-bytes they want and are familiar with the code/data in question so they don't need lots of formatting output or command help, etc. I know mdb engineers have said "file bugs, submit rfe's and we (may/will?) implement them" but those efforts over the years still have only (in my opinion) made mdb marginally more 'user friendly'. Just look at what /net/mage.central/opt/SUNWscat/bin/scat tells you on start (machine name, time of panic, cpu's & mem, os version, etc all things support needs to know up front) Look at the 'help' command how it breaks things down by types of command verses ::dcmds how it lists the stuff alphabetically. The analyze command quickly helps you find typical problem areas while mdb offers no such command. You need to know how to find problem areas using mdb. The way solariscat is built on a shell even makes it easier than using ! in mdb. etc... I'm sure there's a lot more to it than my thoughts above...but it would be nice if mdb could have a bit more user friendlyness built into it going forward. Perhaps the problem with mdb vs otehr tools regarding user friendlyness is that it's dependant on ones experience and requirements from the tool. Later, Kevin.
