On 11/16/18 4:43 AM, Cathy Crumbley wrote: > So, how do we know how to be most helpful for users?
Basically, we don't. Bugzilla tells us about what users expect to be able to do, but are unable to do. Surveys give us what users think that the documentation should cover (^6). Analysing AskBot/web forum data tells what users want to do right now. Analyzing Amazon book sales data tells what information people are willing to pay for. Analyzing TPB/Kat etc. data simply tells us what information people are looking for, not necessarilly what they are willing to pay for. None of that provides us with the information that people don't know that they don't know. By way of example, BAILS is implemented within LibO, (ignoring that part of the functionality was eliminated in either 6.1.2 or 6.1.1). This is great example of something that would be useful in maintaining compliance with SOX, GDPR, and related legislation. Yet the absence of documentation on how to use it, means that nobody asks questions about it. A second example is the built-in version of LOGO --- LibreLogo. There are two or three videos in English, and either one or two short papers, describing a very basic program. There is a long paper in Hungarian, which I don't read, and both Google Translate and Bing Translate mangled beyond comprehension.(^1) What is needed here, is a document that provides both the vocabulary and syntax recognised by LibreLogo. Ideally, a rewrite of K&R, except for LibreLogo, rather than C.(^2) ### I recently read a book on writing Agile Documentation. The main thrust was: * Create four or five user personas; * Write documentation that covers each use-case for each of those personas; Each persona has a different set of use-cases. By way of example: * Molly, the receptionist, whose primary purpose is to look pretty.(^3) Figure out what she uses the product for, and write instructions just for those tasks; * Charlie, the programmer. Figure out what he uses the product for, and writing instructions just for those tasks; * Jake, the accountant. Figure out what he uses the product for, and write instructions just for those tasks; If having trouble figuring out what the individual would use the product for, shadow them for a day or two, writing down when, and what they use the product, or its functional equivalent, for. As an example of use-case specific documentation, look at Thomas Quirk's books on using Excel: * Excel 2016 for Bilogical and Life Science Statistics; * Excel 2016 for Business Statistics; * Excel 2016 for Educational and Psychological Statistics; * Excel 2016 for Engineering Statistics; * Excel 2016 for Environmental Statistics; * Excel 2016 for Human Resource Management Statistics; * Excel 2016 for Marketing Statistics; * Excel 2016 for Physical Science Statistics; * Excel 2016 for Social Science Statistics; The major difference between the books, is the scenarios used in describing when, and how to use the specific Excel function. ### Between built-in functions, included extensions, and third party extensions, Calc has almost 2,500 functions. I recently stumbled across a book _1,000 Built-in Functions of Excel_ It lists a function, the expected parameters of the function, and how the function is used. It seems to me that this is the type of documentation that would be relatively easy for somebody who the time to write, and be very useful to end-users seeking to close the "knowledge that we do not know that we do not know" gap. Distribute it as a book, not a set of pages on the wiki. If they are currently described on the wiki, then transcribe to a book format. If not on the wiki, then add each function to the wiki, as it is being written for the book.(^5) ### In updating _OOo in a Multi-lingual Environment_ for AOo, EO, and LibO. I'm finding that things that could be done with OOo 1.1.3-ZA can't be done with LibO 6.1.3. OTOH, some of the things I couldn't do with that version of OOo, I can do with LibO 6.1.3. I'm documentating those differences in a separate set of documents. The bigger challenge I'm facing, is sticking with a 5,000 word limit for each chapter. (^7) One chapter is one language that uses one writing system, in one country. For Turkish in Turkey, that means four chapters, one for each writing system that has been used in the last 120 years. For Afrikaans in South Africa, that means two chapters --- one for writing Afrikaans using the Latin writing system, and one for writing Afrikaans using the Arabic writing system.(^4) I keep going back to * Should I expect the reader to read the previous chapters; * Assume the only chapter that will be looked at, is the chapter on the country/language/writing system in question; I'm writing for a niche audience, that probably will read the book, before shunting it aside as a reference guide. (Let's face it, how many people can read English written in the Desseret Writing System today? How many people even knew that Afrikaans was written in the Arabic writing system.) On the flip side, it is the type of book that people get, because the title looks intriguing. ##### ^1: Both translate.bing and translate.google provided the equivalent of converting "out of sight, out of mind" to "invisible, insane", in a mammer reminiscent of toki pona's phrase for "ice coffee" --- "solid water hot black bitter drink" . ^2: On second thoughts, I'm not sure I want to see a compiler written in LibreLogo. On the gripping hand, that could make for a very useful Easter Egg. On the fourth hand, K&R is probably the best written book on any programming language one will ever encounter. It contains everything one needs to write a good program in C, without leaving any aspect of the language out. ^3: Most jobs, especially at the lower levels, have a primary function that is radically different from the job description. WallMart's greeters,for example, have a primary function of loss prevention. ^4: It doesn't help that Afrikaans uses glyphs from the Arabic Writing System that aren't part of Unicode. ^5: As I am writing this response, my Internet connection went down. ^6: More than one company has done a multi-million dollar launch for a product that multiple surveys, and focus groups said are highly desirable, only to end up in a crash and burn scenario. ^7: In academia, a reference book is roughly 120 pages, or 10 chapters, or 50,000 words long. Thus, 5,000 words per chapter. jonathon -- To unsubscribe e-mail to: [email protected] Problems? https://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/ Posting guidelines + more: https://wiki.documentfoundation.org/Netiquette List archive: https://listarchives.libreoffice.org/global/documentation/ Privacy Policy: https://www.documentfoundation.org/privacy
