jwcolby
jwcolby at colbyconsulting.com
Fri Oct 16 07:16:46 CDT 2009
It also depends on many other things. For example if you were to convert one of my projects you would be facing my framework and tons of classes and modules. That stuff might just convert but maybe not. How about DAO specific code in modules? I reference databases, recordsets, fields, even properties of fields, I use dynamic queries modifying TDFs in code, I build dynamic SQL and execute it with the database object. I have error logging, login logging, usage of forms logging. I control access to forms and controls on forms based on user/group/usergroup security, manipulating Enabled, Locked and Visible properties based on the user rights. I control subform loading based on whether a ton of controls two forms up have valid data in them, allowing supervisors to override but not normal users. Also what BE are you going to? SQL Server? Staying with an MDB? How about VBA functions embedded in queries? Are you moving from DAO bound to ADO unbound? Are you setting a DAO recordset to the form's recordset clone and then doing stuff behind the scenes? I am now using pass through queries to pull data from SQL Server views. How about import / export of data? My applications strip files off of email and download from FTP sites, and even watch directories, then convert to / from CSV and fixed width files, importing and exporting to tables. I have table driven code which builds fixed width reports with header and footer information based on mainframe specs provided by banks and insurance companies. Your formula would probably work if the original was written by a power user (or summer intern in the case of NorthWind). I would be VERY hesitant to make a firm quote on such a rule of thumb. John W. Colby www.ColbyConsulting.com Shamil Salakhetdinov wrote: > Hello Dan, > > I'd count it as one form/report => 1 day => times 2.5 = that would include > all project costs not only development. > But if you do not plan to change any functionality then it can be done > quicker - here is a real life conversion case's stats online: > > http://northwind.codeplex.com/Thread/View.aspx?ThreadId=54924 > > The estimations were between 45 and 165 hours - just for conversion > development work - average estimation was exactly around the final figure: > ~65 hours for 35 forms/reports (with subforms/sureports). Final result is > not "industry strong" - it would need another 60 hours to make it "industry > strong", even more time would be needed to write docs and online help if > needed etc. - all in all it will be 1 day for 1 form/report for such a > simple application... > > Of course it also depends on how experienced your developers are, what > architecture your converted system is planned to have etc. > > Thank you. > > -- > Shamil