Chuck, Bravo and Amen!
[mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of CRPence
Sent: Thursday, July 10, 2008 7:23 PM
Subject: Re: Search OS/400 Physical Files
Dave Odom wrote:
Seems like the native OS "find" capabilities are WAYYY outdated and
not anything like the Windoz Search function. And it also sounds
like one must depend on utilities created for other OS's or some
other vendor's tools to accomplish a simple global function like
search global for a string which is native in most other OS's. If
true, it doesn't bode well for the IBM i. How sad.
Out from under the Norse bridge, more snide commentary instead of any useful input on how to solve the issue presented by the OP. How sad.
<sarcasm> Way outdated.... So, so true. People are asking week after week for such a function. This forum, USENET, and elsewhere; they are just deluged with such requests over the years. </sarcasm>
If it were such a /simple/ function there would have been plenty of offers for resolution already, by the many who have needed it and coded up the correspondingly /simple/ solution. The issue is extremely complex because the requirements from one individual to another are always wildly different, and as an object-based system instead of the typical /everything is a file/ architecture, the complexity grows in defining the scope.
Even if i5/OS provided some basic function that was thought to be generally acceptable, the same concern would occasionally recur, but with variations on the original theme. Most notably would be all the whining about how the search is honoring CCSIDs or not, that it is identifying numeric data as a match to a character string that was provided as input, that it did or did not find both numeric and character for the hex string provided as input, that it was searching exclusively one of only externally described, program described, or source files, that it was searching encapsulated data but not text attributes of the objects, that it was only searching external objects, or that it was not searching QTEMP of other jobs -- but the list goes on ad infinitum. So really it is probably best left to vendor tools that can cater to some more specific desires, rather than the OS development trying to create something to appease the majority; only to find that it is a generally derided feature because it is so slow & cumbersome because it is too generic in its capabilities, or just that it is not exactly what they wanted [so they had to build or buy what they wanted anyhow].
I am sure if someone came up with a real solid idea of what was considered generally desirable, and there was significant support from something like COMMON with regard to that explicit design concept, then IBM would be willing to provide it. At least then the development would have an idea of what goal to achieve for the customers instead of having to decide what they think they should give to the customers; there is no end to the complaints that all software development is guilty of the latter. But as can be inferred from my sarcasm above, there has in my experience, never been much of a calling for any such function, because what few requests are published have typically been easily solved by an existing [and yes, often non-native client-based] features.
Perhaps someone could start by diagramming what might be a representative command interface to do what might be considered the most commonly required function. I am sure it would soon become obvious just how /simple/ it is when more than one person provides input.
This mailing list archive is Copyright 1997-2015 by MIDRANGE dot COM and David Gibbs as a compilation work. Use of the archive is restricted to research of a business or technical nature. Any other uses are prohibited. Full details are available here. If you have questions about this, please contact