Subject Re: indexing problem From Mervyn Bick <invalid@invalid.invalid> Date Tue, 19 Jul 2016 09:49:56 +0200 Newsgroups dbase.getting-started On 18-Jul-16 11:07 PM, Charlie wrote: > Hi Mervyn... OK I'm starting to get it thanks. > > I am confused though.  You have either with cindex1 or cindex2 and then the parameter is cindexname.  I don't understand that part of it.  How did it go from cindex1 or 2 to cindexname? > > BTW I think I was confused on what the bootstrap is. > cIndex1 and cIndex2 are two variables in your XDML program.  Different names because you have two different tags. :-) In point of fact you don't even need to use variables. You could simply hard-code the tag names as literal values.        if nForm = 1           DO myForm.wfm with "whatever"        else           DO myForm.wfm with "other"        endif It's a personal preference but I seldom use literal values directly in code.  I usually assign the values to variables or user-defined properties of an object. The values after the WITH, be they literals or variables or even properties such as form.entryfield1.value, are the "parameters" that are passed to your OODML form. At the form's end of the transaction  "parameter cIndexName" does two things.  Firstly it tells the form that it can expect a parameter and secondly it gives a variable in which the value passed to it is to be stored.  As only one tag can be passed at a time I used cIndexName for the variable name as this helps me remember what is in it.  It also is meaningful to me as I am going to assign it to the indexName property of the form's rowset. parameter x and f.queryname.rowset.indexName = x would have worked just as well. You can name variables anything you like today.  Back when I started programming (using a VERY basic BASIC on an ICL mainframe computer) all that was allowed were the single letters a to j.  One needed to add copious remarks detailing what was in each variable so as not to get things mixed up in the code. I still use single letters, usually n or i, for counting loops but for the rest I've adopted the convention of starting variable names with a lowercase letter that indicates the type of variable.  c for character, n for numeric, d for date, b for boolean (logical), a for array and so on.  The second letter is uppercase to improve readability.   Many programmers use this convention as it helps to make code "self-documenting" but it is just a convention.  You don't have to use it if you don't want to. The code for a form in OODML is a class definition based on one of the base classes in dBASE.  It is not executable code on its own.  In other words this is just a blueprint for a form.  As it stands it can do nothing.  One can, however, use it to create as many identical forms as one needs. To display the form dBASE needs to be told to create a form object in memory.  This object has an open() method (and a readmodal() method) which it has inherited from the base class which will display the form on the screen. The form designer places this code, which is known as the bootstrap code, in a form immediately before the class definition. parameter bModal local f f = new MyFormForm() if (bModal)        f.mdi = false // ensure not MDI        f.readModal() else        f.open() endif This is actual executable code which can be executed by the dBASE DO command.  Without it you would have to load the class definition into memory, use the NEW command to create an object from the class definition and then open the form yourself every time you wanted to display the form.  One can, of course, still do this.  The little example form I posted the other day for launching a report to be saved as an HTML page in a file with the date as the name does exactly this. Mervyn.