FilePro Questions?

[Scott Walker]

Brain,

Thanks for your thoughts.  I was really just looking for License Manager
issues that may have been resolved or at least might be resolved if fP was
offered without it.  But I value your opinion on other issues.  Have there
been other commercial products that have gone Open Source and changed their
revenue model from making money on software licenses to support.  I believe
that is where the money really is.  Give away the software to get the
customer and make money on support, customization, training, etc.  The
transition, however, could easily put most companies out of business, since
one revenue stream totally dries up while the other one has not yet ramped
up...better have a good year or two of cash on hand to make it to the pot of
gold.

Regards,

Scott


> -----Original Message-----
> From: filepro-list-
> bounces+scottwalker=ramsystemscorp.com at lists.celestial.com
> [mailto:filepro-list-
> bounces+scottwalker=ramsystemscorp.com at lists.celestial.com] On Behalf Of
> brian at aljex.com
> Sent: Tuesday, March 20, 2012 5:42 PM
> To: filepro-list at lists.celestial.com
> Subject: RE: FilePro Questions?
> 
> Yeah, I'm off-by-one ... 5.0.14 is ok, 5.0.15 has the manager.
> 
> The problems we had are both technical and principal. Actually I don't see
> a distinction between "ideological" and "technical" in this case.
> 
> I would say it's absolutely a hard technical problem that I simply do not
> want to voluntarily start depending on something that we lack the power to
> ensure it continues to work ourselves. It's been proven a falsehood so
> many times over it's not funny whenever any entity makes any kind of
> promises or assurances about their continued existence, terms, etc. So
> relying on that is a guaranteed shot in the head some day. Guaranteed. So,
> it's just completely stupid to voluntarily sign up for it. It doesn't
> matter if they're nice guys or not or what their current or even past
> performance is. If fp were even remotely like _any other_ language then it
> might not be that big of a deal because it would be possible to migrate to
> some other similar thing some day if need arose, but as many have found
> out the hard way, it's practically impossible to migrate off of filepro.
> All you can do is start over utterly from scratch and generally lose
> something along the way even if you gain other thin  gs.
> 
> We don't want to steal anything. We have a successful business and the
> trouble we could get in isn't even close to worth it. We would love to
> give fp lots of money if they would give us actual fixes and features we
> asked for.
> 
> Ask Barry about this absolutely heinous retarded brute force hammer
> process we have to run all the time on some boxes just to keep them from
> having this mystery lockup problem that locks everyone on the entire box
> when it happens. It has to scan through about 120 filepro files,
> multiplied by N qualifiers, 9 times per minute, all day every day, and run
> several commands for each of those iterations, including ps & w snapshots
> and an expanded version of lockinfo to run showlock on each individual key
> file and analyze the results against the process tree and find the process
> associated with each locked record and do some further logic to determine
> which of those process is causing the magic problem locking everyone else
> up. This isn't normal record locking contention. This is something that
> causes every other invocation of clerk on the box to halt when it tries to
> display any record in any qualifier in any file. Until Barry did rather a
> lot of work to track it down as far as he d  id, we couldn't even clear
> the problem any way but rebooting the server. At least this rediculous
> brute force high overhead process clears the roadblock without having to
> reboot mutiple times per week, sometimes multiple times per day. We
> installed strace wrapper scripts around all the clerk & report binaries
> and made the biggest most-paying customers suffer slowness for weeks just
> to try to capture clues when the problem strikes, and only got so far
> without access to the source. Theories and guesses only. If we could just
> insert a few simple, _simple_, lines of debugging we could probably make
> this whole problem go away forever, either by fixing it or at least
> knowing why it happens and how to avoid it. We _tried_ to pay for support.
> Ken worked on it some but for some reason that was never made clear to me,
> it just never went anywhere. I can't imagine we weren't doing our part
> with paying or supplying files or access to real systems. I wasn't
> managing that job but I know we
>   were all in agreement it was a high priority problem that was worth
> whatever fp wanted to fix it. Just to be clear this lockup problem is not
> one caused by the license manager, it's a problem we have now on 5.0.14
> and had at least on 5.0.13 and unknown versions before that. The reason I
> bring it up is it illustrates that "I need the source" or at least "I need
> to know I can at least backup/restore/migrate forever on my own" is not a
> "merely" ideological stance. It's a real technical problem. Voluntarily
> paying for work and support is fine. Being actually dependant on someone
> elses graces and continued existence, offerings, terms, for your continued
> existence is a guaranteed shot in the head some day. It's stupid to
> voluntarily sign up for that.
> 
> When we tried to go to 5.0.15 both to get some of the 5.0.15 features as
> well as in hopes it might resolve the lockup problem, it did not fix the
> lockup problem and caused a few new problems that we frankly didn't bother
> to track down. We already have full time jobs to do and it shouldn't be
> our job to pay even more, for an even more limited product, and then have
> to do _FP's_ work debugging obscure problems that are intermittent and
> would take a lot of expensive man-hours to track down, all the while
> having to apologize to _our_ paying customers for crashes and buggy
> behavior. I can't go into detail about the problems because we never
> figured them out ourselves before we said "Screw this. We'll use 5.0.14
> since at least it's "the devil we know", and just start working on an
> entirely new codebase using other languages even if it takes the next 10
> years."
> 
> Well we do still use fp but it's been several years since then and we have
> a lot of new stuff that only uses fp as a replaceable back end. This has
> both reduced our need for fp licenses while still growing user counts, as
> well as inch us towards the day we are no longer at the mercy of something
> inscrutable and intractable like filepro.
> 
> I still wish more people would have been willing to collect together an
> offer to buy filepro with the express purpose of open sourcing it. If FP
> was Microsoft or Oracle you could reasonably say "The terms may change and
> I may be forced to buy things I shouldn't need to repeatedly, but at least
> MS will always exist and will always be large enough to guarantee that I
> never have un-fixable problems, so it's not a stupid business plan to rely
> on them." But that's not the case at all. It IS stupid to pin your whole
> life on this product and this company. Having so many years of code
> invested in filepro that it's impossible to switch to anything else means
> to me that it's worth almost _anything_ to get the source, and further to
> re-license it so that it's impossible to ever lose it.
> It's worth _more_ GPL'd than it is to try to sell protected copies of it
> to the pathetically few people who would ever even want it. GPL'd, it
> gains immortality and at least a _chance_ of free improvements over time.
> It also gains transparency. No more mystery what the hell is it doing in
> there. It gains at least a _chance_ of new people deciding they want to
> use it and thus replacing some of the lost talent and community. There is
> _no_ chance of this as it stands. None. It's only getting harder and
> harder to sell a filepro based product every day, including yourself as a
> programmer not just the code you've written already.
> 
> If we (the remaining dependant community) bought and open sourced filepro,
> what about the poor fp employees? They're all gone by now _anyways_. We
> (Aljex) would probably gladly employ Ken ourselves if nothing else and who
> else is there anymore? Doing what? Nothing too hard to do somewhere else.
> 
> If it were open sourced and Ken opened up shop himself as paid fp support
> and customization programming that was at YOUR direction instead of BUD's
> he'd probably have work for life at twice or more his current salary and
> on his own schedule and terms. Hardly a hardship. And YOU/WE would get
> more out of it than we get now.
> 
> It's basically priceless to be able to pay a programmer to work on open
> source code. You get so much more bang for your buck. The best few
> thousand I ever spent was to get a tiny handful of small changes made to
> PuTTY. Even though I had to pay 3 different developers for some of the
> same features because the first two did crappy jobs or stopped
> communicating and/or never delivered source before getting a guy who was
> not even really a Windows programmer but was at least better at C than me,
> and understood the goals clearly and was good to work with. The final
> result was worth it all because now I have it, it works _exactly_ the way
> I want, and it's completely unencumbered. I can stick it anywhere and make
> the most effortless user experiences because there is no licensing or
> legal worries, only technical ones, and all technical problems are
> solvable. It's open ended forever now. Every new user install reduces that
> initial investment to insignificance.
> 
> Conversely every time a box reboots or any change to /etc/hosts changes
> the way a box resolves itself, facetwin breaks and we have to wait until
> at least 9am central to get it fixed, which is about 6 hours too late
> after many of the earlier east coast users need to start working, And we
> pay $110 to $150 per _seat_ for this! vs a one time investment of a few
> thousand and then free forever and no such thing as license broken ever.
> It's just no contest at all.
> 
> I'm sorry I get so frustrated over this topic because it's so obvious to
> me that I get incoherent when I allow myself to get sucked into commenting
> at all. But you can't just only ever sigh and walk away. Once in a while
> you do have to try to say why you feel a certain way.
> 
> I wanted to answer Scott's question but about what specific problems we
> had but I can't unfortunately because as I said we decided given
> everything else, it was a poor investment of resources to debug them. We
> had a "bad but good enough" hack to brute force 5.0.14 into working and so
> we decided it just makes more sense to just live with 5.0.14 and invest in
> our own stuff instead of investing in fp and just to ice the cake even
> have to apologize to and even lose customers while we did so.
> 
> That was only 5.0.15, so absolutely, years later by now 5.7 may not have
> those problems whatever they were, or, even if they still exist, your code
> may not tickle whatever problems our code tickled. They do at least try to
> be more accommodating than most companies by offering several different
> methods for implementing the license manager so if you don't like
> chocolate flavored poison, you can have caramel or strawberry instead.
> 
> --
> bkw
> 
> 
> -----Original Message-----
> From: "George Simon" <flowersoft at compuserve.com>
> Sent: Tuesday, March 20, 2012 10:50am
> To: brian at aljex.com, filepro-list at lists.celestial.com
> Subject: RE: FilePro Questions?
> 
> I'm running 5.0.14 and it doesn't have the license manager.  At least the
> version I have, does not have it.  I think 5.0.15 is the first version
> with
> it.
> Like you, the license manager is stopping me from upgrading a boat load of
> customers to 5.6 or 5.7.
> Just not worth the aggravation.  If something goes wrong with the license
> manager the customers are going to be calling me, not fpTech, to solve
> their
> problem.
> 
> -----Original Message-----
> From: filepro-list-bounces+flowersoft=compuserve.com at lists.celestial.com
> [mailto:filepro-list-
> bounces+flowersoft=compuserve.com at lists.celestial.com]
> On Behalf Of brian at aljex.com
> Sent: Monday, March 19, 2012 8:35 PM
> To: filepro-list at lists.celestial.com
> Subject: Re: FilePro Questions?
> 
> 5.0.13 is the last version without the license manager and the last
> version
> we will run. We tried newer versions and the manager injected more and
> worse
> problems than any of the improvements took away. As much as I'd like some
> of
> the features, the license manager makes it a non-starter.
> 
> --
> bkw
> 
> -----Original Message-----
> From: "Jay Ashworth" <jra at baylink.com>
> Sent: Monday, March 19, 2012 7:53pm
> To: filepro-list at lists.celestial.com
> Subject: Re: FilePro Questions?
> 
> ----- Original Message -----
> > From: melancon at cajun.net
> 
> > Gulf Island QC is having problems with losing network connection and
> > FilePro not handling it too well. Had some missing or corrupt data.
> 
> Yup; database apps are singularly intolerant of network problems,
> "reliable
> transport" or no.
> 
> > IT claims the version of FilePro we are running is old and needs to be
> > upgraded.
> 
> Maybe, but not for this reason.
> 
> > I feel it may be hardware such as a bad switch that could possibly
> > causing the problem. Some user are having more problems than others.
> 
> You are almost certainly correct.
> 
> Download WinMTR and run it on each workstation against the IP of the
> fileserver, for, say, 5-10 minutes per machine.  Any problems should be
> immediately apparent.
> 
> > Would upgrade to FilePro 5.7 help improve network communications?
> 
> No.
> 
> > Is my 5.0 not capable of dealing with newer servers?
> 
> There are some very uncommon circumstances where you might see specific
> network apps have trouble with newer file servers, but if your *client*
> OSs
> are up to date as well, it's maybe .01% or less.
> 
> > Should I upgrade to 5.7?
> 
> Now, this is a much more nuanced question.
> 
> Since, unlike Ken, I am *not* employed by FPtech, I can point out that
> 5.0.14 is probably the last version *before* all the crappy stupid license
> manager garbage got included.  [Checks]  No, I'm wrong; 5.0.14
> *has* it, so there's probably no reason except money not to upgrade.
> 
> > Can I still use 5.0, now that Window 7 is here?
> 
> I'm relatively certain that answer's yes as well.
> 
> > I appreciate any feedback? I truly feel there is nothing wrong with my
> > 5.0 version and IT is just making excuses.......
> 
> Off the top of my head, I agree with you, but you failed to mention one
> very
> important fact for the diagnosis:
> 
> Was your network *functioning well on all the OSs and hardware it's
> currently on, and then went downhill*?
> 
> Or did something get upgraded, and *then* everything went in the toilet?
> 
> Cheers,
> -- jra
> --
> Jay R. Ashworth                  Baylink
> jra at baylink.com
> Designer                     The Things I Think                       RFC
> 2100
> Ashworth & Associates     http://baylink.pitas.com         2000 Land Rover
> DII
> St Petersburg FL USA      http://photo.imageinc.us             +1 727 647
> 1274
> _______________________________________________
> Filepro-list mailing list
> Filepro-list at lists.celestial.com
> Subscribe/Unsubscribe/Subscription Changes
> http://mailman.celestial.com/mailman/listinfo/filepro-list
> 
> 
> 
> 
> 
> _______________________________________________
> Filepro-list mailing list
> Filepro-list at lists.celestial.com
> Subscribe/Unsubscribe/Subscription Changes
> http://mailman.celestial.com/mailman/listinfo/filepro-list
> 
> 
> 
> 
> 
> 
> _______________________________________________
> Filepro-list mailing list
> Filepro-list at lists.celestial.com
> Subscribe/Unsubscribe/Subscription Changes
> http://mailman.celestial.com/mailman/listinfo/filepro-list