WHY

H

Harlan Grove

linear regression?

it allows you to model 128 different dimensions
....

How? Simple pairwise correlations? Do you know what linear modeling is?
if you don't believe me when i say that I have a billion records with a
sub-second response time-- you should have told me this a lot earlier.

IT HAPPENS USING OLAP. IT HAPPENS A THOUSAND TIMES PER DAY ACROSS THE US.

If it's cached, it's fast.

Reality check. Most machines have at most 2GB physical RAM. Nontrivial
records span at least 100 bytes each. That doesn't include indices. A
billion nontrivial records take up 100GB. Last time I checked, 100GB doesn't
fit inside 2GB. If it's not all in RAM, I don't accept your assertion of sub
second response time. So I don't accept your assertion of subsecond response
time for billions of records - at least not on the machines most people
outside IT departments have. And if you mean the OLAP data and processing is
hosted on some server farm, maybe if you have one for your exclusive use you
may have this sort of response time, but not when there are a few hundred
other users and the server has to balance processing across several tasks.

So if you're running all your favorite software across all the 20 machines
in your own home, your results may differ from real world results.
OLAP TOOK AWAY ALL OF THE PERFORMANCE PROBLEMS OUT OF DATABASES

BY CACHING! It's a function of how much data can be stored in memory. The
more you can store, the faster your system will appear to be.
with OLAP recordcout is irrelevent.

To be replaced by dimension count.

You must not understand how CPUs work. If you have 1,000,000 records in a
single field table stored on a RAM disk (no longer a Windows concept - a
nonpersistent file system using system RAM as the storage medium), and an
OLAP cube with a single dimension and 1,000,000 entries along that
dimension, any CPU will take about the same time totaling one as the other.
If there are N numbers to be summed, the CPU needs to perform N-1 ADD
operations. If OLAP cubes appear to total arbitrary numbers of records
almost instantly it's because they totalled them when they were fed from
tables into the OLAP cube and cached for later use. A useful feature, but
not what you think it is.
 
H

Harlan Grove

where did you get those numbers from-- A DATABASE HUH?

Whatcha think? Where have I stated that databases aren't useful for what
they do best?
http://www.netcraft.com/Survey/Reports/current/graphs.html

This report isn't very current.. all I know is that Microsoft came out with
Windows 2003 Web Edition with a base price of about $399 like 18 months ago

Not what I used. I made 30 separate queries for individual domain
information. For example,

http://uptime.netcraft.com/up/graph/?host=www.microsoft.com
and it is easily the fastest webserver i've ever seen

Given the extremely limited perspective you've already proven time & time
again, what you've ever seen is likely to be an extremely small subset of
web servers.
it runs circles around apache; it runs circles around Win2k

Uh, bozo, Apache isn't an OS. Do you mean the latest version of IIS running
under Windows Server 2003?
Solaris-- one of the largest Operating Systems from that survey--- Sun
only has a what $5bn market share? Do you really expect them to be around
forever?? LoL

Do you mean market capitalization? You really do need to learn some basic
business terminology. It might make you come off as a grown-up at least some
of the time.

If Sun open sources Solaris, which they've been spreading rumors about doing
for years, then, yes, it'll be around for decades. Software lifetime doesn't
depend on hardware lifetime. As a hardware company, I'd agree Sun is toast.
But IIS compression in Win2k03 is AWESOME

http://www.port80software.com/surveys/top1000compression/

22% decrease in html size?

That isn't something to sneeze at.. It'll take over; and it already has

It may (or may not), but you're being an even bigger fool to believe it's
happened already. More deluded inaccuracies from you. Whatsa madda, you
miffed because I proved you knew squat about one aspect of the *current*
state of the internet?
 
H

Harlan Grove


More gems of rhetoric from you.
VB is the most popular language in the world

As I've pointed out before, if you define 'language' to include functional
programming languages with bui;t-in grid controls, there are at least 20
times more Excel 'programmers' than VB ones.

And since there are millions of programmers *NOT* writing code on or for
Windows, VB may have the plurality of developers, but no where near a
majority.
The fact that Firefox and opera don't allow client-side vbScript has
nothing to do with security.. . . .

Yes it does. Client-side scripting without some form of execution control
facility is necessarily insecure.
just because VB is popular; and it has some vulnerabilities-- that doesn't
mean that it's a bad choice for client-side scripting.

**ANY** vulnerabilities makes it a bad choice, bozo.

And it's not the language itself that's evil, it's client-side scripting
without execution control that's bad. VB[Script] is an irrelevance.
ActiveX downloads in IE might be a bad idea.

No, ya think?!
But vbScript is innocent and very powerful

Java BS is for wimps

JVMs are secure. Case closed. [It's the difference between Unix-originated
software which was written by people with a clue about network security vs
Windows-originated software written by people who couldn't break through the
conceptual barriers of one user per machine.]
 
H

Harlan Grove

Microsoft is losing the database war because of this marketing problem with
Access.

Simplistic.

They can't win on non-Windows systems for obvious reasons.

As for Windows systems, it's a question of who's generating the reports. If
it's the IT department, then I'd imagine Crystal Reports is usually
preferred to Access. If outside of IT departments, then Access is often
overkill. If the needed data resides on a server, any ODBC client can fetch
the data. If a lot of manual entries are included in the report,
spreadsheets are more efficient.

The day of the single PC DBMS ended back in the mid 1990s. As soon as
networks became ubiquitous in most larger businesses, the benefits of
centralization were recognized, and databases returned to the tender care of
IT departments.

As for the department in which I work, the people who populate the company's
databases do most of their data entry into character mode applications
originally written in Clipper. Data is collected on local file servers in
..DBF files which are automatically replicated to home office in the wee
hours every night, and are subject to IT validity checking before being fed
in batch into mainframe databases. Any sensible person's reaction to this
configuration should be if it ain't broke, don't fix it.

No spreadsheets in this process. No Access either.

As for what I do, mostly it's statistics and discounted cashflow analysis in
which the timing of cashflows varies from project to project. That means a
lot of data entry with little likelihood of reusing those entries. That may
affect my perspective on the efficiency of data reuse - it just doesn't
matter to me, and I suspect most other moderate to heavy spreadsheet users
face the same situation. And since most of the data I use comes from sources
outside my company, there's very little I can pull from in-house databases.
Access would be more of a hindrance than a help for me. I realize you just
can't understand this, but I've stated it anyway.
the marketing problem with Access is due to beancounters like you who think
that it is 'too difficult' to have to save the file before you start typing

I have to save XLS files before I can do much useful with them.

No, the problem with Access is that it doesn't provide me with any benefits
in what I do. To repeat, I don't get much of the data I need from in-house
systems, and that which I do I can fetch from within Excel. I don't need
Access's query builder, I'm comfortable writing my own SQL. I don't have any
use for .MDB files. I have no use for Access forms since I perfer entry
straight into spreadsheets (free form entry into a grid beats dialog-based
entry into separate fields any time for me, but I'll grant that's a personal
preference and subjective). And I don't produce reports per se, so Access's
reporting abilities do nothing for me either. And if you think it's easier
to structure ANY output exhibit using Access's Report tool than Excel,
you've never had to generate the sorts of exhibits I have.
Microsoft just needs to get their head straight.

No, they see clearly that Access will never be a major revenue source.
Beleive me, if it could be, Microsoft would be pushing it like you wouldn't
believe. You're confused because you rely on it so heavily you believe
everyone else should do so too. Well, the rest of the world doesn't need to
think like you. (THANK GOD!)
The continued marketing problem with Access is smearing SQL Server

They're separate products. SQL Server's problems lie in the fact that,
despite what you believe, it's not as well suited to many database tasks as
mainframes, and it may not be as easy to administer the Windows systems on
which it needs to run as it is to adminsited many *x systems.

You're also ignoring vendor lock-in, which Microsoft has made so much use of
itself. Office is predominant now becasue Lotus completely screwed up in the
early 1990s by believing Windows was no big deal. Microsoft made them pay
for that. It was different with databases. It wasn't mostly non-IT people
who'd be using them, and IT shops already had DBMSs they were using. It's
not as easy as you may believe to change from one DBMS to another, so the
transition would be gradual at best. Access would have minimal influence on
this.

Besides, as I've mentioned before, most IT departments don't give non-IT
users database access beyond what's needed to pull data into client
applications. Access isn't needed for that.
and SQL Server can and DOES compete extraordinarily well against those
LEGACY (obsolete) databases.

Not if those other systems are already paid for. And you're only trotting
out your ignorance if you believe other DBMSs haven't been continuously
upgraded. They may not have all the stuff Microsoft has, but many IT
departments may have decided that all that extra stuff isn't needed or is
less useful than you believe it is. To an APL programmer, there's no other
piece of software that's anywhere near as efficient for number crunching,
but only a fool would believe APL was a mass market system. Perhaps ditto
for OLAP/MDX etc.
Microsoft is just failing to make marketshare gains like they need to.

Failure to make marketshare means buyers don't believe the benefits of
switching to MSFT stuff outweighs the costs. How does this square with your
assertions that MSFT stuff has brought about a revolution? Maybe you're the
only one who believes that?
and it is a marketing problem, not a software problem

So? Microsoft has *NEVER* been shy about marketing anything it thinks will
make it LOTS of money. Perhaps they believe Access, SQL Server, OLAP/MDX
won't make them much money, certainly not enough to justify a big marketing
push. Unlike you, I'm not so arrogant that I think I know how to run MSFT
better than they do themselves. But arrogance and militant ignorance seem to
be at the core of your being.
they need to either embrace Access or drop it.

You don't understand ROI. If it makes decent returns given minimal marketing
costs, they'd be foolish to drop it. But you're the champion of foolish
things, so not surprising you're advocating foolishness.
when I mean embrace Access they need to do this:

1) convince beancoutners like you to get on with your life and start using a
real application

Clever. I use it more than anyone else in my department, but only when it
makes sense to use it. Not for things I can do more efficiently using other
software. And REMEMBER, I don't write reports.
2) fix bugs in Access

Why should Access be special? Microsoft is slow to fix bugs in all the
Office apps.
3) tell us where Acccess is going to be 10 years from now

Dead? An add-on for Excel?

As a simple query creation tool it's unnecessary MS Query is sufficient.
There are similar tools now on Linux systems. QBE wasn't novel when Access
1.0 rolled out, Paradox had provide it for 6 years previously. Access also
wasn't the first PC-based DBMS to provide SQL, R:Base 3 was.

As a table manager it's unnecessary. Most DBMSs come with GUI DBA tools.

As a report generation tool it's unnecessary. Crystal Reports is at least as
heavily used.

It's a convenience tool, not a necessary tool.
4) make more people have access; include it with a standard edition of
office

More stupidity.

To repeat, Microsoft know a LOT better than you how to make money. They're
not about to include it into the standard edition when they can charge more
for it in professional.
 
H

Harlan Grove

Adobe Acrobat is the biggest BLOATWARE i've ever seen.

do you really need to check for updates every fuc#ing day adobe?

do you really need to be a TSR (terminate and stay resident)

?

Acrobat is useful and it isn't Access. That must be why you don't like it.

It serves a purpose. If you don't like Acrobat, there are other PDF creators
and readers. If you don't like the concept of PDF files, you're an idiot.
Well, you're already an idiot, and it seems there's no limit to how big an
idiot you are, so this is just more evidence of what everyone knows about
you rather than any new revelation.

Few sensible people make use of Acrobat features beyond those provided in
version 3. No need to check regularly for updates. If it's become a service
(so a full-time memory hog) in recent versions, revert to version 4 or 5.
Access has included this thing called the snapshot viewer that comes with
Office; and it allows you to do this without BUYING ANOTHER PRODUCT

Yup. Access just does everything better. Soon you'll tell us how Access is
the tool to use to create computer animated movies.
I'll bet you download the Flash viewer too huh?

?

You do find it difficult to stick to thinking and logical argument.
 
H

Harlan Grove

it isn't licensed; it is a CA basic IDE and compiler.

has nothing to do with Microsoft.

I wrote BASIC on my commodore 20 years ago.. i didnt' need a license from
Microsoft for that, did I?

So just the VB syntax. What about the GUI development tools, or do you just
mean the language? If just the language, get used to the fact that BASIC is
unloved outside Windows. Mainframes had VS BASIC. Macs have RealBasic. Linux
and BSD (if not others) have XBasic (which also has a Windows port). They're
cult languages at most on their respective systems.

On Windows, VB has its place because of the quality of its GUI construction
tools in versions 1 through 3 (Windows 3.1 days). It had nothing to do with
the language aside (probably) from Microsoft recognizing that BASIC was so
little used outside Windows that VB would provide more OS lock-in than any
other language. Recall that TurboPascal had more developers on MS-DOS
systems than QuickBasic (or BASICA for serious applications) prior to VB's
debut.

And if you could write a BASIC interpreter 20 years ago, what stops you from
adding a VBScript module to FireFox yourself?
 
H

Harlan Grove

it would be GROSSLY inefficient to do sorting on 100 different machines

It happens all the time. For example, my wife, my syster, her husband, my
father and me all have each other's phone numbers in each of our cell
phones' memory. Each of these cell phones presents an alphabetized list of
stored numbers. Ergo, each of these phones has had to sort the same numbers
on separate machines.

IDIOT! Next thing you'll spew forth is that no one needs anything more than
a thin, dumb client connected to a single hypercomputer to avoid having to
store any datum or sort any list more than once.
AND THEN YOU HIT THE 64k LIMIT AND YOUR SCREWED

If I'm working with large datasets, it'd be in R or some form of math
software. I use Excel when I have small datasets. So not a problem.
you can sort in a query; on the server side.. and it is a LOT more powerful
than sorting in excel

Sorting in queries is INEFFICIENT when the data to be sorted isn't in a
table or an OLAP cube but in an array in memory having just been calculated
from other data in memory. Thoroughly idiotic (so thoroughly you!) to
imagine it'd make sense to put it into a table in order to sort it.

Example: real time techincal analysis for stock trading, looking at moving
averages of stock prices from transaction to transaction. The data may be on
it's way to a database, but it'd be read and used in calculations prior to
reaching the database. If the program were ranking changes in such moving
averages, guess what? They won't be in a database table. They'll be in an
array in memory. And since this information would need to be used
immediately after the next transaction comes through, it'd be absurdly
foolish (so you'd think about doing it) to store it in a table. So how would
you sort this sort of data? In memory.

A different sort of application with similar processing needs would be real
time chemical concentration monitoring systems. Their readings may
EVENTUALLY be stored in database tables, but processing uses values in
memory in simple and efficient data structures for processing, not in DBMS
tables.

You don't comprehend this sort of application, so not surprising you can't
conceive of how they'd work and what routines would be needed.
if all the numbers come FROM the database; why are you worried about needing
to push them back into the db in order to sort them?

The number DON'T come from databases. Is this a sign you might be coming to
understand this? Or are you assuming some fantasy world in which all the
data one ever needs to work with comes from an in-house database?
if you leave them in the db and sort them there; its a lot more manageable

True and irrelevant. If they're NOT in any database to start with, and
they'd be used outside the database once sorted, it's stupid to pull them
into the database just to sort them and push them out afterwards.

Can you conceive of some people using computers to analyze data that doesn't
come from databases?
 
K

Ken Wright

I think my biggest bugbear with databases is the people that use them, and the
reliance that some of these people put on the machine being inetlligent enough
to do things that they intended to ask instead of what they really asked.
Drives me absolutely nuts when I ask someone for a report on something and they
give it me and I can immediately pick holes in it, even though it isn't my data
but theirs.

Recent case was when I asked for a set of Actuals (Costs) on a particular
program broken out into the various cost buckets that we use. The 'analyst'
gave me the report, and after about 5 seconds I look at one of the buckets and
told him it couldn't possibly be right because this bucket of cost appeared to
be circa 60% of the Total sell Price, when in fact I know that that cost element
should be only a fraction of that. His answer - "Well I ran the query and
that's what the report says" - Grrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrr

I take the report over to someone else that used to work there and asked her
"what's wrong this, only I don't understand how this could come out" - Instant
response - "Aaahhh, but you can't use that figure because it's the total for the
whole cost account and not just that program"

The first 'analyst' has no real concept of what is in the data, and because it
is a database believes that what the query says must be right!!! Second analyst
always knew her stuff, and was always right up there in my estimation.

Too many people believe they can simply push a button and take what comes out -
My respect goes to those that will then trawl through and look for those little
red flags that jump out when something just doesn't look/feel right. Little
checksums/benchmarks etc that when applied to a set of data can point to
underlying issues that may not be readily apparent to the majority of people.
Far too many people accept what they see instead of questioning it, simply
because "it's in the database".
 
A

aaron.kempf

harlan, harlan, harlan

so you can recreate the same report in excel week in and week out

and then you can RUN it if you need to.

The thing is that with Access, you can schedule with a batch file
whatever you want to do.

Can you do that in Excel?

HAHAHAHAHA grow up kids

excel is a dead-end
 
A

aaron.kempf

you dipshit; ive handled billions of records with a sub-second response
times on SQL Server.

SQL Server took over the world many years ago; but some companies are
too diseased with Excel and DB2 and Oracle in order to modernize.

Grow up guys; excel is dead. Oracle is dead.

Unix is dead.

SQL Server dual-processor machines cost what $15k with licensing and
hardware?

Unix can't compete on price or performance
 
A

aaron.kempf

Ken

im so sorry that you had a bad experience with a database professional.

technically, database people are overworked and spreadsheet people jack
off and write the same report every month.

hire more, better database poeple.
put all of your beancounters thru Access and VB training.
 
A

aaron.kempf

you guys try to use excel as a one-size fits-all solution

except excel doesn't scale to 66k rows and it sux because it is SLOW
and can't be trusted.

it isn't stable.

it's just spinning your wheels.. instead of redoing the same report
every week; do it right ONCE and then have a batch file print your
report.

excel is over-used and it is a disease
 
H

hrlngrv

(e-mail address removed) wrote...
you guys try to use excel as a one-size fits-all solution

Untrue.

You say we do, but it's just a straw man you have to return to because
you're unwilling to or incapable of rational discussion of using
anything other than DBMSs.
except excel doesn't scale to 66k rows and it sux because it is SLOW
and can't be trusted.

Slowness is relative and needs to be placed in context. If I need to
calculate some value (e.g., acceptable bond portfolio as collateral in
lieu of a letter of credit), I may need 100-odd entries no where in my
own company's databases in order to calculate one figure as result.
Since the data entry would be the same for both spreadsheet and
database approaches, the formulas or queries would represent the entire
difference in modeling time. I don't know about you, but for me this
would be MUCH MORE QUICKLY done in Excel than Access.
it isn't stable.

Depends on what's being modeled and how it's being modeled. There are a
large variety of general discounted cashflow models that vary only in
terms of discount rates and cashflow term and pattern. When such models
are designed correctly (and even weekly cashflows over a span of 100
years requires fewer than 5500 rows), the formulas don't change. Only
the inputs change, and (to repeat from above) new data entry
requirements would be identical for spreadsheet and database.

There are thousands of other types of models that rely on static
mathematical development. Only the extent of the data may change from
use to use, and it's possible to write formulas (or, better, defined
names) that can adapt themselves to the extent of entered data.

Misuse of spreadsheets as databases can cause problems, and I'm not
sure anyone in this thread has disagreed with that point.
it's just spinning your wheels.. instead of redoing the same report
every week; do it right ONCE and then have a batch file print your
report.

If generating simple, periodic reports that could be fed directly from
databases were the only thing for which spreadsheets were used, you'd
have a point. However, whether you're aware of it or not (and whether
you possess the minimal intellect necessary to learn and appreciate
this or not), this isn't the only thing for which spreadsheets are
used, and most of thos other tasks would be cumbersome to implement in
databases.
excel is over-used and it is a disease

It may be overused, but you can't count as overuse tasks performed in
Excel by people without access to any other programmable tools. You can
bemoan the lack of access to DBMSs, but it's a cost-benefit trade-off,
and you're quite obviously incapable of appreciating that.

The only thing diseased in this thread is your mind, but I suppose
atrophy would be more accurate.
 
H

hrlngrv

(e-mail address removed) wrote...
so you can recreate the same report in excel week in and week out

Understood.

You're too stupid to understand what I write.

You're too stupid to be able to imagine using any software for any
purpose other than writing reports or you're irremediably deluded into
believing any & all computer use results in reports.

You're too stupid not to need to return to demolishing your straw man
argument about recreating reports in spreadsheets.
The thing is that with Access, you can schedule with a batch file
whatever you want to do.

Ditto for Excel or any other scriptable application development tool.
Whether most people do rather than repeating manual steps every time
they need to repeat some task is irrelevant with regard to how I use
the tools I have at hand. If you want to argue with the hypothetical
spreadsheet users who spend their workdays cutting & pasting from one
report to another, wake up and realize that they don't read this
newsgroup.
Can you do that in Excel?

Yup. Apparently you can't. Sad, really.
 
K

Ken Wright

The only dead end is the path your logic takes. It falls over with almost
every new post you make. Every time Harlan posts a logical counter to one
of your rants that you can't answer, you have to retort with an insult
because you know you got beat in war of words. Like I said before, you were
checkmated on your opening move, but you just weren't bright enough to see
it.
 
A

aaron.kempf

technically harlan

i was watching some ESPN show the other day; and they were talking
about coaches being able to consume tons and tons of videos-- and i
swear to god i saw NFL coaches sitting there; wading thru hundreds of
movie clips using-- guess what-- Microsoft Access. A nice lil
application with forms and a couple of wizards.. it was obviously
written in MDB or ADP I coudnt' tell.

if it is good enough to run your company off of; and it is good enough
for the NFL coaches-- why the hell isn't it good enough to automate
your reports, loser?

re: pdf

pdf sux ass

it is a TSR-- those engineers and the whole Adobe Company should be
SHOT for making a Terminate and Stay Resident Program.

I believe that is in the same category as SPYWARE

who knows what it is doing-- you open a PDF and you close Acrobat---
and it still hogs memory.

Those people shoudl be ****ing shot for letting that out on the
marketplace; that is called screwing with my computer performance. And
I won't put up with it.

For you LOSER EXCEL PEOPLE Microsoft came out with a PDF killer back
for Office 97, it is called the 'snapshot viewer'

It allows you to export a report into a small file (much smaller than
PDF) and emial it around making it portable.

Where MS screwed the pooch was in not including this in Excel. Now
Microsoft tries to push a new 'PDF Killer' onto us-- and im like '****
you microsoft; i've already got the snapshot viewer; why would i want
to deal with this 'Microsoft Document Imaging' bullshit.

Adobe Sux
Microsoft can't compete

and Access fits my needs-- it has a PDF viewer included called the
'Snapshot Viewer (.SNP extension)
 
A

aaron.kempf

RE:
If just the language, get used to the fact that BASIC is
unloved outside Windows. Mainframes had VS BASIC. Macs have RealBasic.
Linux
and BSD (if not others) have XBasic (which also has a Windows port).
They're
cult languages at most on their respective systems.


So BASIC is unloved; except for the FACT that it is the worlds most
popular language on the worlds most popular platform.

And didn't you just illustrate how EVERYONE LOVES BASIC?

If BASIC wasn't loved outside of the Windows world; why would
mainframes have VS BASIC? Why would macs have RealBasic? Why would
Linux have XBasic

It is the world's most popular language..

You don't need to learn VB/VBA you need to learn MDX harlan or else
you're roadkill.
 
A

aaron.kempf

Sorting in queries is INEFFICIENT when the data to be sorted isn't in a
table or an OLAP cube but in an array in memory having just been
calculated
from other data in memory. Thoroughly idiotic (so thoroughly you!) to
imagine it'd make sense to put it into a table in order to sort it

there is no such thing as an array.

there is only databases.

anyone tries to tell you otherwise and they're tryting to sell you
something

why would you reinvent the wheel?
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Top