[DUG] Validating CDS files

Matthew Comb matt at ferndigital.com
Tue Jan 18 13:42:09 NZDT 2011


Bravo!

> I'm wondering if you analyzed your need for popcorn carefullly. You asked
> for popcorn.. you got your popcorn. but perhaps you've been taken for a
> ride. If we analyze your need carefully for an hour, what we could have
> determined is that you truley wanted a 30 minute comedy show, and some
> salt
> - and -vinegar chips to go with it. Not popcorn at all. Perhaps the desire
> for popcorn stemmed from your not correctly seeing your needs in the first
> place. I feel its my duty to help you access the situtation so that you
> get
> what you need and not what you want.
>
> We could discuss this further privately, and I'm sure with all my
> experience
> in these situations that a 3 hour lunch break daily, with said included
> comedies and chips will satisfy your needs more appropriately.. The
> Popcorn
> is just a bandaid.
>
> On Tue, Jan 18, 2011 at 12:44 PM, David Brennan
> <dugdavid at dbsolutions.co.nz>wrote:
>
>> Mmmm, I'm enjoying my popcorn as I watch...
>>
>> -----Original Message-----
>> From: delphi-bounces at delphi.org.nz [mailto:delphi-bounces at delphi.org.nz]
>> On
>> Behalf Of Matthew Comb
>> Sent: Tuesday, 18 January 2011 11:12 a.m.
>> To: NZ Borland Developers Group - Delphi List
>> Subject: Re: [DUG] Validating CDS files
>>
>> > Those developers that came back with "TO DO" lists were perhaps
>> themselves
>> > not taking the time to understand the customers problems.  The
>> customer
>> > told
>> > them "those 10,000 features are great, but we'd really like X, Y and
>> Z",
>> > so
>> > the developers simply came back to you with that list of "X, Y and Z".
>> >
>> > Far from analysing too much, they were most likely not analysing
>> *enough*.
>>
>> What your argument does not take into account yet again (which was my
>> point that was missed), is the time and money factor. If you have a 1
>> day
>> or 2 day install, you do not have time to spend all 2 days taking on
>> customer requests when the customer is not yet up to speed with the
>> capabilities of the system. A customer often doesn't know what they need
>> until they get to use the system, then their requirements frequently
>> change. If you have unlimited time and money then by all means, spend
>> every waking moment analysing each and every item raised, Cameron would
>> probably tear his hair out.
>>
>> >
>> >
>> > That specialist you hired is most likely looking at the X, Y and Z and
>> > then
>> > asking the customer... "But what do you need X, Y and Z for?" then
>> showing
>> > them that they already HAVE X, Y and Z among the 10,000 other
>> features.
>> >
>> >
>> > That quite neatly demonstrates my point I think.  Understand the
>> > problem...
>> > not only might the solution you are asking for not be ideal, but you
>> might
>> > already be in possession of the solution !!!
>> >
>>
>> I don't think it does at all.
>>
>> >
>> >
>> >   "I have a legacy piece of software which very very rarely corrupts a
>> >    local data file - fact"
>> >
>> > The real problem to be solved is how/why the corruption occurs...
>> > everything
>> > else is just sticking plaster, not a solution, and even the sticking
>> > plaster
>> > can only help if you apply the right sort of dressing correctly.
>>
>> Thats right, that is the root problem, but as this piece of software has
>> been in the field for 8 years, and is being decommissioned, there isn't
>> a
>> lot of point in that is there. So a bandaid is more than adequate.
>>
>> >
>> >
>> >   "Regardless of the better long term solutions, we will take
>> >    a defensive programming measure to protect against the loading
>> >    of corrupt files."
>> >
>> > And my point in this area was that unless you understand how/where the
>> > corruption is occurring, the defensive measures you might take may
>> very
>> > well
>> > end up not being defensive at all.
>>
>> If you validate a CDS file, and don't load it if it is corrupt. That is
>> a
>> defensive measure.
>>
>> >
>> > e.g. the idea of checksuming the output from the server...  if the
>> data
>> > coming from the server is corrupt from the get-go, then check-summing
>> > helps
>> > you not one little bit.
>> >
>>
>> Agreed which is why I didn't ask about it.
>>
>> >
>> >
>> > You initially blamed the corruption on a flaky wifi connection -
>> highly
>> > unlikely for the reasons that I and others have explained.  You then
>> > explained that you were streaming the data, so the corruption might
>> occur
>> > in
>> > sporadic bytes in the stream.
>>
>> I do not believe I did suggest the contents of the stream would be
>> changed. What I said was that a TCP steam can be disconnected.  this was
>> yet another wild assumption on your part. If you send a 1GB file through
>> a
>> stream, it is NOT as Kyley suggested an All or Nothing approach. You are
>> taking bytes into a stream. It is up to you whether or not you discard
>> the
>> stream based on a disconnection or not. That decision point is out of my
>> control as it is Midas + DBX4MySQL which manages this.
>>
>> >
>> > But then you seemed attracted to the idea of using XML as a potential
>> way
>> > forward, which would seem to contradict the streaming nature of the
>> data
>> > transmission.
>>
>> The incorrect assumption you made at the beginnning was that the CDS
>> file
>> was coming from the server as a file payload. Its not. The client is
>> connecting to MySQL DB server directly. All suggestions and
>> misunderstandings about hashing, file downloads, etc were based on that
>> incorrect assumption.
>>
>> >
>> > And now you admit:
>> >
>> >    "I do not know yet where this corruption occurs. DB Server,
>> Drivers,
>> >     In Memory, Saving of the file, Server DB."
>>
>> I don't, and it doesn't matter. Why ? Because at a rate of 1/10000 it is
>> a
>> manageable issue, and with the software being decommissioned it is good
>> business sense to investigate the server mechanics. its better to
>> bandaid
>> and move on.
>>
>> >
>> >
>> >
>> >
>> >
>> >   "Its not true to take the standpoint that if someone asks for the
>> > solution
>> >    then they should be able to provide it e.g  If I ask for Word 2010
>> > should
>> >    I be capable of sitting down and writing the application suite?"
>> >
>> > I didn't say that they *should* be able to provide it in the sense
>> that
>> > you
>> > took it to mean - I said as to draw attention to the inherent
>> > contradiction.
>> > The only way you can confidently ask for a solution in the form of a
>> > solution specification, without explaining the problem, is if you
>> fully
>> > understand the problem and *could* provide the solution yourself.  In
>> > which
>> > case, you would never have needed to ask someone else to provide the
>> > solution for you.
>>
>> Thats just rubbish.
>>
>> >
>> > In other words, when someone asks for a solution, rather than stating
>> a
>> > problem, the likelihood is that they are asking for the wrong solution
>> > from
>> > the very start!
>>
>> Statistically probably correct more often than not, but still an
>> arrogant
>> assumption that the person is clueless / uninformed of possible
>> solutions.
>>
>> Its also arrogant to assume you know everything abou the motivation and
>> drivers as to why the person asked the question in the first place.
>>
>> For example in this case, the fact that the software is being
>> decommissioned and the fact that its rarity does not make a server
>> solution fiscally viable.
>>
>> >
>> >
>> > Your "Word 2010" example this is a quite different sort of request
>> from
>> > what
>> > we are actually discussing (it's not a "technical solution" just a
>> > shopping
>> > list), never-the-less even this inappropriate parallel can be used to
>> > illustrate the same point:
>> >
>> >
>> > If you ask for Word 2010, should I just sell it to you simply because
>> you
>> > can't write it yourself ?
>>
>> That logic much like the rest of your logic is flawed. I did not say
>> that
>> you should sell it to me because I could not program it. You should sell
>> it to me because I asked to buy it.
>>
>> I am asking to buy it, because I don't want to program it.
>>
>> >
>> > Or should I first ask what you need it for.... ?
>> >
>> > If you tell me you want it for doing bulk emails, and you heard that
>> Word
>> > 2010 has some neat mail-merging capabilities, then I am pretty
>> confident
>> > that I can sell you a better solution for your needs than just handing
>> > over
>> > a copy of Word 2010.
>>
>> Again you are incorrectly assuming that the person asking the question
>> is
>> a retard.
>>
>> >
>> > I could just give you what you want, but that may not be what you
>> need.
>> >
>>
>> Its not your decision to make.
>>
>> >
>> > Incidentally, nobody "gave me" a blog.  I created my blog in order to
>> > share
>> > my skills and experience and - if my stats are to be believed - many,
>> many
>> > people appreciate it.
>>
>> Actually I've found your blog to be very good and useful. I feel that if
>> you reigned in the arrogance a tad, you might get places faster.
>>
>> > I gain some sense of satisfaction from the fact that people (well,
>> *some*
>> > people at least) seem to appreciate my effort, but it certainly
>> doesn't
>> > "go
>> > to my head".
>> >
>> > _______________________________________________
>> > NZ Borland Developers Group - Delphi mailing list
>> > Post: delphi at delphi.org.nz
>> > Admin: http://delphi.org.nz/mailman/listinfo/delphi
>> > Unsubscribe: send an email to delphi-request at delphi.org.nz with
>> Subject:
>> > unsubscribe
>> >
>>
>>
>> _______________________________________________
>> NZ Borland Developers Group - Delphi mailing list
>> Post: delphi at delphi.org.nz
>> Admin: http://delphi.org.nz/mailman/listinfo/delphi
>> Unsubscribe: send an email to delphi-request at delphi.org.nz with Subject:
>> unsubscribe
>>
>> _______________________________________________
>> NZ Borland Developers Group - Delphi mailing list
>> Post: delphi at delphi.org.nz
>> Admin: http://delphi.org.nz/mailman/listinfo/delphi
>> Unsubscribe: send an email to delphi-request at delphi.org.nz with Subject:
>> unsubscribe
>>
>
>
>
> --
> Kyley Harris
> Harris Software
> +64-21-671-821
> _______________________________________________
> NZ Borland Developers Group - Delphi mailing list
> Post: delphi at delphi.org.nz
> Admin: http://delphi.org.nz/mailman/listinfo/delphi
> Unsubscribe: send an email to delphi-request at delphi.org.nz with Subject:
> unsubscribe




More information about the Delphi mailing list