Right now, I'm reviewing some datasheets of parts that might be used in future projects, and I am just appalled at how little proofreading was done.

I have the datasheet for an ADC that, according to the revision history, was revised twice. It's only about 40 pages long, yet both of the "revised" versions I have come with errors in table references (they say "See table X", when the information is actually in table Y and table X contains completely and obviously unrelated information). It contains timing diagrams that, to my understanding, violate constraints that are later explained several times in quite some detail - "Do not clock out data for for X ns before and after event Y", yet in the timing diagram, event Y happens somewhere smack dab in the middle of the data transfer.

Also, I have the datasheet of an MCU. In the chapter about a certain peripheral, it does not give a lot of detail and instead states "See the actual MCUs datasheet for details". Huh ? I thought that is what I'm looking at. There is no other datasheet. Ah well. Maybe that's why the thing is labeled "Preliminary". So when do I get the actual finished thing ? (Note to Erik: Yes, that is from the company whose MCUs you don't use. I don't want to call any names here.)

Any comments about incomplete, confusing and/or self-contradicting documentation ? Hints on how to deal with it ? I've found out that support from the manufacturer can be quite minimal if you're not buying their chips in the millions (i.e. you're not making cellphones or similar mass products).