It has been confirmed that OI32 and NT Service 1.5 will work OK together provided that OI32 doesn't feed NT Service 1.5 any records ]64k for writing.

In our case it is very unlikely that a data file record would exceed 64k - and anyway an MFS could guard against that (Sprezz have one available via their web site).

However, my concern is Btree records in the !file.

In OI16 (and Arev) the Btree index algorithm clearly splits and balances Btree index records so that none exceeds 64k.

Does OI32 use the same algorithm as OI16? If not then OI32 writing Btree records via NT Service 1.5 could cause some fun.

Can we have an authoritative response on this point?

Thanks in anticipation

Having produced more documentation on index structures than any other organisation we'd hope this can be considered authoritative.

Index row sizes in AREV/OI16 and OI32 are controlled by a common area that dictates that the maximum node size can be 1K. Same goes for transactions.

Speed improvements may be gained by increasing this node size (see next SENL) but this is quite a technical tweak so most systems have a 1K node size.

In short this won't be a problem in OI32.

The Sprezzatura Group

[i]Celebrate CeBIT with Sprezz Local![/i]

Authoritatively, this is not a problem in OI32. We still use the same routines that break the nodes into 1K chunks.

Thanks-

Mike RUane

Many thanks, Mike & Andrew.

I'm sure I won't be the only person asking this question!

View this thread on the forum...