All this talk about DataBases

[Andre Garzia]

On our lexicon stack we have each card name being the word it stores,  
we use a combination of filters and RegEx to search the data, this  
way, I don't need to loop the cards, I just need to iterate over the  
cardnames, thats why it is fast. I think if you create smart indexes  
and use clever regular expressions to find stuff in this index,  
revolution performs very well.

so in the end, we have our storage in a stack database, but our  
search routines use indexes to find stuff. Still everything is on RAM  
and RAM is faster than disk access.

andre



On May 30, 2007, at 1:06 AM, J. Landman Gay wrote:

> Scott Kane wrote:
>> I'd be curious to know why RR decided to change the behaviour of  
>> how stacks are read (from file as opposed to loaded fully into RAM).
>
> They didn't; the engine has always worked that way since its  
> original MetaCard incarnation. Scott Raney, the creator, wanted  
> speed and so wrote the engine to load everything into RAM. The  
> trade-off is that you need as much RAM as the size of your stack --  
> which is one reason why a dedicated database is more suitable for  
> huge data sets.
>
> Raney's standard suggestion was that databases larger than 5,000  
> cards should be moved to a "real" external database. Stacks under  
> that number perform acceptably well. I have pushed it to 10,000 or  
> so without any particular indexing or special efforts, but it does  
> slow down, especially if you use the "find" command to locate  
> content. HyperCard had its wickedly fast "hint bits" search that  
> Rev doesn't have. When Rev searches, it has to look through the  
> text of every field which is much slower. However, if you create an  
> index and the scripts reference cards by ID (the fastest way) you  
> can increase the number of cards without too much lag.
>
> I wrote a database with over 40,000 records, and for that one I  
> loaded a text file into RAM and then used a 1-card display stack to  
> show the desired record. This method requires that you write all  
> your own navigation and search handlers, but it was about as fast  
> as HC when I was done.
>
> -- 
> Jacqueline Landman Gay         |     jacque at hyperactivesw.com
> HyperActive Software           |     http://www.hyperactivesw.com
> _______________________________________________
> use-revolution mailing list
> use-revolution at lists.runrev.com
> Please visit this url to subscribe, unsubscribe and manage your  
> subscription preferences:
> http://lists.runrev.com/mailman/listinfo/use-revolution