OK, I see now you weren't just making a database for the keys to your cars in the back-40. :-)
It looks interesting, and I would like to spend a little time thinking about it, but I have a couple of questions just-for-now. (See interspersed remarks) Ralph Shumaker wrote: >.. > OK, making things a little less simple, how about this: > > Bear with me as I attempt to present this in a way that doesn't mangle > it all into a big mess. > > The H75 appears to have *zero* substitutes. But that kee is used by > several Ford vehicles (not to mention Mercury, Lincoln, and others). > Just Ford alone has 25 entries using that kee. > > It's not as simple as saying "the Ranger takes the H75" because the H75 > was used by the Ranger from 97 thru 98 and during 05. Some (not all) > models of the Ranger used the H75 in 96, and 99 thru 00. And the Ranger > only accounts for 4 of the 25 entries for that kee. > > If I'm making a kee for a vehicle that calls for the H67 and I'm out of > that kee, then I can substitute the H54, H55, H62, or H66. Now, I'm not > sure I trust the H55 and H62 info, because only one car lists those two Where does the list (lists) come from? Car mfgr? Kee (keemaker?) mfgr? > as substitutes. Other than this one car, all but 2 of the others list > both the H54 and H66 as substitutes. Two cars list only the H54 as an > acceptable substitute, but probably the H66 will work too (in other > words, the exclusion of the H66 on those two lines was probably a typo). > > All substitutes are nearly identical to the first kee in the tail of the > kee. Most differences in substitutes are in the head or neck of the > kee. The H62 has a rather large head and probably could not substitute > for any other kee. On some, the neck is not as long and would prevent > them from substituting in for other kees. The H54 can substitute in for > several others (including the H62) because it has a long neck and a > relatively small head. > > There are 25 entries that call for the H67. One of these lists the H54, > H62, and H55 as substitutes. Two list only the H54 as a substitute. > The other 22 list the H54 and the H66 as substitutes. > > For my concerns, I don't care that one car lists different alternates > than all the others. My main concern is that each primary just show a > list of possible alternates. Maybe I will add a field to show how many > times a particular alternate is listed for a specific primary. But I > don't feel the need to track which car calls for which alternates. The > book I get the information from often has information changing from one book: Some kind of 3rd party compendium? A trade tool, maybe. > year to the next, and *not* necessarily for corrections. I think this > book is manually typed each year... heh! people still do that? =-O >.. Information that is correct one year > can be wrong the next. I only want my database to show which primary > kee the car calls for. And my database will have info from different > years of the books to help compensate for the errors. I may even add > another field to indicate U, C, or W for Unconfirmed, confirmed Correct, > confirmed Wrong, and possibly even V for Varies. Just a thought: it may be much easiest to do this in 2 stages. 1. assume there are no such conflicts to worry about, (but at least reserving the thought there will be) 2. extend (1) to accommodate those conflicts > > On Ford, the H62 was only used on the Escort, and even there, only from > 1991 thru 1996. The H54 is the only listed substitute. > > Only Ford (not including transponder kee entries): > > Entries Primary Substitutes > 2 H70 > 25 H75 > 22 H67 H54 H66 > 10 H51 H53 > 36 H50 H52 > 2 H78 > 2 H71 > 1 H67 H54 H62 H55 > 4 H54 H60 > 1 H76 > 1 H62 H54 > 1 H53 H51 > 1 H52 H50 > 2 H59 MZ16 > 2 H65 MZ27 > 2 H66 H54 > 1 S1186TS > 10 H54 H60 H67 > 1 1185T-P H54 > 1 S1185T-P H50 > 1 H55 H54 > 2 RV4 > 2 62FT > 1 WB2 > 2 MZ5 MZ10 > 1 MZ9 > 1 MZ4 > 2 H67 H54 > > Or, put another way: > Primary Substitutes > H50 H52*36 > H51 H53*10 > H52 H50*1 > H53 H51*1 > H54 H60*14 > H54 H67*10 > H55 H54*1 > H59 MZ16*2 > H62 H54*1 > H65 MZ27*2 > H66 H54*2 > H67 H54*24 > H67 H55*1 > H67 H62*1 > H67 H66*22 > MZ5 MZ10*2 > 1185T-P H54*1 > S1185T-P H50*1 > > Now, the ones that only show 1 occurrence in this final list, I'm not so > sure that I trust. For example, the only occurrence of the H52 being > replaced by the H50 is more likely to have been a typo in the book, and > that car *actually* call for the H50 and replaceable by the H52. Some > of the other probable typos are less obvious. > > Essentially what I want to do is to make my own automobile kee reference > manual, first in database form, then printed. I want to create a > database that contains all the information from all the manuals that I > have. Basically, each line from the manual in my hand will be > identified in my database as having come from the 2007 manual... the 2007 manual? the 2007 issue of the trade book mentioned above? >.. I will > probably end up duplicating each line and correcting the info, and call > the new line my own. And when I print it out, I will print only *my* > lines. > > When entering the data, I don't want to have to type in 1184FD/H54 (the > full description of the kee) each time and increase the chance of > introducing typos (which is obviously how the makers of this manual are > doing it). And seeing as how the H50 is listed as the kee of choice in > 30 different entries in Ford alone, I think that kees would be a good > candidate for their own lookup table. > > When it comes down to it, I want to have a table where I've already > entered all the kees, another table with all the transponder info, transponder info comes from another reference? Is this uniquely determined by car, maybe? (make, model, year, [sub-series]) > another table with all the substitutes, another table with all the > notes, another with car makers, another with car models, another with > lock applications, and another with code series, all enterable with a > combo box, one that will show possible matches based upon the first few > letters I type, but allowing a new entry if it matches no others. OK the UI is a separate (and significant) question but I like the thought of thinking in advance how the database ought to support your access requirements. > > But the main table is where they are all connected. The main table will > have: > an entry for the year (of the manual) > a combo box for the brand of the manual (currently only 3) > a combo box for the model (currently over 700) > a combo box for the car maker (currently under 100) You don't actually have combo boxes in the database, at least not in relational databases. You're talking about the UI, I believe. > an entry for the year from > an entry for the year thru > a combo box for the lock application (currently under 100) lock application? meaning > a combo box for the code series (currently about 100) series is a disjoint classification of kees, maybe? > a combo box for the kee of choice (currently ???) > a combo box for substitutes (fewer than the kee of choice) > a combo box for transponder info (under 100?) > a combo box for the notes. (under 100?) > > If the car model is filled in first, then most of the time the car maker > will only have one choice, but could have as many as three (currently). > > Most lines have zero substitutes, some have only one, but some have more > than one. Same thing goes for the transponder info and the notes. 0-or-more is easy to do in relational databases. > > Most of the info in the manual is repetitious, but uniquely combined on > each line for a different lock and kee application. For example, the > Ranger has 17 lines, each one unique from the others in some way. In > other words, each field has a small number of possible entries (other > than car models and probably kees). But no two records will be entirely > the same. There are approximately 2,000 records for each manual. And > most of the information that is possible in each of the fields will be > identical from one manual to another. I'm thinking that each field > should have it's own table, except the years (because of how simple they > are, like entering 4 digits should be easier that scrolling thru a list). > > I don't think this is all that simple. But then again, I've never > embarked upon anything quite like it. > Later.. ..jim -- [email protected] http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg
