Canned POC

[Flawless_koder]

Ok - here's the problem.

We had a canned proof of concept for a web application.

The data was held in Javascript arrays and array hashes.

The problem now is that I've had to proove a concept that had to go beyond that.

I used ( as per usual with me ) PostgreSQL as the RDBMS and HTML-Mason ( perl ) as the serverside language.

The SQL database is simple in structure - five tables i think, one sequence and a view.

The HTML-Mason is rather more complex - but i'd hope to think, given the hype everyone's always making over these "new langauges", that one of the more recent langauges could handle the same stuff ( i've got a feeling it's going to have to be asp - you'll see ).
[ It's strange me calling other languages new when Mason itself is - but it's pretty much perl - so i can still be cynical as per usual ]

So - I now need to take this web application that came from a poc, and has now be turned into a web application, and turn it back into a clientside poc.
I don't mean clientside as in "page served by server - all the action happens in scripting langauges inside the html".
I mean clientside as in "the pages themselves will live on a cd - and any action must happen on the demonstration machine with no intervention from a server for anything.

So - my question:

How can I achieve a RDBMS in a "canned" form.
The demo pc can't be expected to be running any sort of RDBMS.
There can be no assumptions as to software on there ( before anyone suggests access [ for which i would have shouted anyway ] ).

So can I achieve the same thing now with ASP and some sort of demonstration sql ?

Has anyone had any experience with this sort of thing?

Thanks

Flawless

[MarcusJT]

Originally posted by Flawless_koder
There can be no assumptions as to software on there ( before anyone suggests access [ for which i would have shouted anyway ] ).

So what browser(s) will this app have to work on? If IE4+, then you can assume that certain things are present.... (I have some ideas, but it all depends on what browser(s) you are targeting)

[Andrew-J2000]

just do this from a flat text-file? not sure what exactly you need it to do, but thats the only non-rdbms that comes to mind straight away?

[Flawless_koder]

No - needs to be faster than text files.
I was hoping for a type of mini Sql system really.

( and so obviously looked at mSql - but i don't think that will do it )

The browser is ie5.5+ ( as with all poc's )


Thanks guys

Flawless

[MarcusJT]

As part of the IE 5.5 installation, a slimmed-down version of ADO 2.1 is installed, which supports Access 97/2000 DBs via ODBC (and possibly via JET4 I think, you'll have to check). The ADO installation is contained within the MDAC_IE5.CAB file.

Therefore, you could easily do anything you usually do in ASP via client-side scripting in VBScript/JScript/JavaScript using ADO and an Access DB of your choosing!

So, I don't think you even need those arrays and hashes (unless they are doing something I haven't understood)!!!


(HTH)

[Flawless_koder]

They are doing something you haven't understood

The postgres database contains in excess of 3000 records.

There are five tables, a sequence and a view.

Can this easily be converted into an access database, do you think Marco ?

Flawless

[MarcusJT]

Originally posted by Flawless_koder
There are five tables, a sequence and a view.

Not being familiar with PostGres, I do not know what a "sequence" or a "view" is in this context - perhaps you could enlighten me.

However, I find it hard to believe that PostGres can store and query data in a way that MS Access cannot replicate directly (or at least simulate)!! And a 3000 record, single table Access DB isn't really a problem, is it?

Of course, when using an Access database on a read-only medium such as a CDROM, you need to make sure that connections are made read-only so that ADO does not attempt to create a locking file.

[Flawless_koder]

You may have to guide me through the rest of the details ( read only etc )

A view is a stored query that creates a kind of table.
Instead of thumbling phrases - i'll let you read it:
http://www.postgresql.org/idocs/inde...ial-views.html

A sequence is what you use to determine the next value in a default value system.

For example - Primary keys use an auto incrementing value system by default - in Postgres creating a primary key automatically creates a sequence for it.
It's a queriable entry that allows you to control the way the incrementation etc happens.

Does this info help?

Flawless

[MarcusJT]

Thumbling?
Do you have a lithp?
(I don't mean to take the pith, Flawleth. Honethtly )

Ok, a view in PostGres = a stored query in Access (which I am currently writing an article about, so no probs there)

As far as sequences go, what specifically are you using it for? Can it be simulated in script, perhaps with the use of another table (as a lookup) in the database?

(apart from that, I don't see any problems porting this to client-side script + ADO + Access so far...)

[Flawless_koder]

I use the sequence to make sure the id's of the five tables are unique between each other - they all use the nextval of the sequence for their default id value.

Flawless

[MarcusJT]

So a given RecordID can only exist once in the database? What is the rationale behind this? Are the tables identical in composition? If so, why not one big table? I'm sure you have your reasons, but at the moment I don't know what they are!!!

[Flawless_koder]

The reason is as follows:

When i use the stored query - i want the id field to be selected and be unique - yet still refer to an individual table and record.

The data in the five tables isn't similar enough to normalize.

Flawless

[MarcusJT]

Presumably you have to a reverse-lookup to find out which table contains a given ID then.... Is that what the hash tables are for?

[MarcusJT]

Discussion moved to ICQ for speed.......