Why Excel is not a stardard format

Hi gang

This is about the boss. He wants a couple of branches to deliver some address data periodically as an Excel file. His argument is that the Excel format is the most widely available de facto standard and thus would reduce costs for the branches.

For me it’s just a big PITA (LAMP here based on openSuSE-11.1). Can you give me some good arguments why Excel is not a standard format and is not well suited to pass data from one organization to another?

It’s not fully documented because it’s proprietary. You don’t want your server app breaking because somebody at the sending end used some proprietary feature. Tell him that this might happen at the most inopportune time, e.g. Friday afternoon.

If he wants to pass spreadsheets, use CSV. It’s a text format, well-specified, supported by spreadsheet programs and there are libraries to parse it for many languages.

I agree with ken_yap.
Another reason to use csv is that you can import csv data not only into spreadsheets, but also into just about any database as well and address info, suggest a database somwhere, even if it’s only in a contact manager or email program.

While I support that csv is the (much) better solution, I know that bosses may not be that intelligent. Maybe the fact that OpenOffice can export as xls helps you. My OpenOffice even offers three (!) different xls formats (so far for it being a standard, ask your boss which one of those ‘standards’ he wants).

And the new version of M$ Office/Excel will probably have a new and different format for .xls again and yet another Excel ‘standard format’ will arrive on the scene.
This will probably also create problems for older windows excel users as well.
If the boss is intent on using excel format, tell him to use win2000 version since this is probably better supported.

I really hate OpenOffice and Excel and all similar applications because they can’t be used from bash scripts. There is no way to automate things when using those apps to generate a certain format.

My data is kept within a MySQL database. I could live with CSV for data exchange, because I can easily generate a CSV-file with a neat and sexy bash/sed script. But the spongeheads at the other end of the pipeline will convert everything again to stuff it into some M$ database. Crazy world.

Another argument I was thinking of: Excel is not a standard, because the format may/will change with every new release coming out. If one of the branches updates its Excel there is no guarantee that center will still be able to use the files sent.

Anyone knows how many different Excel formats there are already in use?

I’m a little bit short of arguments because I’m not familiar with windows. The last time I used it was when it was called “Windows for Workgroups”.

I also like to ask for CSV because I can then write all sorts of shell scripts to manipulate them. The others are always amazed that they can send me a 10^6 lines where a huge percentage contains a certain error and that I send them a corrected CSV in a nick of time. They can not understand that i did NOT clicked and typed all those fields lol!

Note that because you are asked to parse XLS, it’s a harder task. For instance, XLS supports multi-sheet spreadsheets. In OO, it only exports to CSV the current one. So what happens if your user sends you a multi-sheet file? Can you be sure that your parser (which at best would be some kind of converter using OO called through UNO) can handle everything the users might throw at it? Besides that uncertainty, factor in more development time.

> Can you give me some good arguments

will you actually risk your job to win this disagreement with your
boss??

my advice: do the work in OO and then “save as” an excel doc…


platinum

On Tue, 08 Sep 2009 10:26:01 +0000, vodoo wrote:

> I really hate OpenOffice and Excel and all similar applications because
> they can’t be used from bash scripts. There is no way to automate things
> when using those apps to generate a certain format.

There’s a really good series of articles in Linux Format about how to do
this sort of thing with OpenOffice. I don’t recall offhand if they cover
spreadsheets, but ODS spreadsheets are basically zipped files that can be
parsed with a script.

Jim


Jim Henderson
openSUSE Forums Moderator

Fortunately I will NOT loose my job just for discussing data interchange formats with “The Boss ™”. The boss in this case is the government and I am in charge of an NGO. The problem is, that people at the other end know nothing except Word and Excel and they will tell their IT Dept. what the want. If I succeed to make my point that they are going to run into trouble they will hand the problem over to the specialists. Much better for me, because they will understand.

No, I don’t want to use OO. Documents saved with OO in excel format are not 100% compatible with M$ excel. I want to solve the problem, not be part of it rotfl!

  • There are way more VBA macro viruses than OO.org ones.
  • OO.org does not require you to pay for updated versions and once one of the members updates you’ll need to add a converter of some sort in order to READ the data but it may not allow you to WRITE to the data. For example, I’ve been moved to Office 2007 for supporting purposes and probably one of the biggest uses of that is opening files sent to us from partners and clients who already upgraded.
  • CSV is more portable and the benefits of putting it into a database can be significant (depending on data and uses)

Why not just use Google Docs or Zoho for everything, which makes it easy to work together on?

You can get one answer from the horse’s mouth http://office.microsoft.com/en-us/excel/HA103357121033.aspx
though it is buried in the small print.

Excel and Access don’t use the same standard for dates and Excel can create errors in the data during data exchange. There are even different versions of Excel which need to be handled differently.

On Tue, 2009-09-08 at 08:36 +0000, vodoo wrote:
> Hi gang
>
> This is about the boss. He wants a couple of branches to deliver some
> address data periodically as an Excel file. His argument is that the
> Excel format is the most widely available de facto standard and thus
> would reduce costs for the branches.

Unfortunately he’s right. Just like mp3 is the common de facto standard
for music and blu-ray is the de facto physical media for HD.

In all cases, these are proprietary, protected formats that are not
open.

A “de facto standard” does NOT have to be an open one… sad, but true.

Just means that you’re ok with complete and total vendor lock in (you
could argue that the boss is ok with vendor lock in for other things as
well… which probably isn’t true, but something for the boss to
consider).

>
> For me it’s just a big PITA (LAMP here based on openSuSE-11.1). Can you
> give me some good arguments why Excel is not a standard format and is
> not well suited to pass data from one organization to another?

We have an OOo server in batch mode that allows us to convert documents
into different formats.

Anyway, I don’t like what you’re boss is doing, but at the same time,
he’s correct about Microsoft’s proprietary formats being a “de facto
standard” today.

I do not know if it helps, because Zwitserland is not in th EU. But my governmebt (and the EU) are pro the usage of open standards and I would have a strong point in telling them that what he wants is contrary to the rules.

I once succeeded in bringing the webmaster of our city in contact with a bureau of the home office that had a big web-site about do’s and don’ts in governement web-sites (should be W3C in HTML and CSS validated, specialy with the disabled and their special browsers in mind).
She did not even know there were rules for her web-site! A bit of presure helps.

You can lose a lot exporting to CSV from a spreadsheet app, e.g., formulas, charts. By my experience, most bosses need as many pictures as possible to read and comprehend anything more complex than a timesheet :slight_smile:

Hi
I use to use Crystal Reports and then Managers, Accountants etc
could just click on a desktop icon to generate their own customized
reports from all sorts of data sets.


Cheers Malcolm °¿° (Linux Counter #276890)
SUSE Linux Enterprise Desktop 11 (x86_64) Kernel 2.6.27.29-0.1-default
up 1 day 6:55, 3 users, load average: 1.40, 1.24, 1.31
GPU GeForce 8600 GTS Silent - Driver Version: 190.18

Hi

Thank you all who gave some input. It was very helpful. My position has now been written and sent to “The Boss”.

Feel free to continue the discussion if you like.

What about Pentaho BI? Heard anything about it?

Do you know how to program in Java? I’ve written a lot of java code that parses existing xls spreadsheets and creates new ones. They can be as simple or complicated as you wish and it supports multiple sheets, etc. You just need the jawin api (or poi) with java.

If the spreadsheets will always keep the same format, this would be an automated way to handle it.