>> So for me it's hard to understand why to reinvent >> it, if we already have it by the name "RDD". >> >> If we want to go beyond this and implement the whole >> "ORM" buzzword, it's a whole separate project, >> not just some extra lines of code in sqlmix. > > ORM is defined as Object Relational Mapping at wikipedia, but there is > another meaning for that, that is Object Role Modeling acording to > www.orm.net. > Maybe the example pointed by Massimo is more related to Object Persistence. > Wikipedia moved/renamed Object Persistence page to System Prevalence > (http://en.wikipedia.org/wiki/System_Prevalence) and acording to it, the > advantage of this aproach is "Simply
Wow :) Actually I wanted to find the intended meaning of "ORM" in recent discussions based on context (but had a hard time), and above perfectly shows the whole confusion above these three-letter buzzwords. > keeping system state in memory in its normal, natural, language-specific > format is both orders of magnitude faster and more programmer-friendly than > the multiple conversions that are needed when it is stored and retrieved from > a DBMS." > One case of use is that objects can be collected in an array, so once they > are loaded it is much faster to access ascan form them than using searching > in a disk file. Maybe I'm wrong, but above description seems to cover the database access concept implemented in all dBase-like languages. Except that the "array" is present on RDD level, so you don't really make separate disk accesses by each FIELD->x reference (but I'm not 100% sure on that in current Harbour implementation). So, if that's all "ORM" means, we already have it, and that was one of the features that made dBase so popular versus other "database" languages. Brgds, Viktor _______________________________________________ Harbour mailing list (attachment size limit: 40KB) Harbour@harbour-project.org http://lists.harbour-project.org/mailman/listinfo/harbour
