SAP proves, yet again, that Excel is utterly unkillable How unkillable is Microsoft Excel? SAP has just shown it might be nigh-on immortal because despite having built an analytics cloud it claims is capable of end-to-end decision-making, it’s decided it also needs Office 365 integration. “Many businesses tell us that one of the biggest challenges with their data is having no …
IMHO there is no other application that comes close to Excel in terms of usability, versatility and usefulness.
Disclaimer: No other application has been harmed in the creation of this message.
Disclosure: I have no financial or other interest in Microsoft or other software vendors
people who think it's a database program
OK. Mini rant here.
This doesn't apply to big databases that definitely need the full relational hoo hah. But most don't.
<rant>In my youth there were numerous flat form databases. Simple, easy to use. And for most purposes they did all you needed. e.g. List of members' or clients' names, date they used the service, date of payment etc etc. With some basic search facility. How many clients joined in January, that sort of thing.
Simple, easy to use. Limited learning curve. Didn't need a specialist to set up. Then they all seemed to vanish.
Now there's just Access, which is much more complicated to set up and if you're not used to it, not at all intuitive; or stuff that's even scarier/more expensive to buy/expensive to set up/complicated to use than even that.</rant>
But there's also Excel.
Which pretty much does what people need it to do for most use cases. So they start using it. And once they start on it they stay on it.
I think maybe there needs to be a set of DB tools that can scale from flatform. Starting simple and easy, but with guidance to make itself into something a bit more complex, one step a a time as a need grows
Not disputing that Excel can do the job, hell I use it myself a lot when working with CSVs for data input to PowerShell scripts etc. I've even used it on rare occasions to produce rack diagrams as the cells map nicely to rack units, even though Visio is a much better (albeit much more expensive) tool for the job.
The problem (IME) is that in my long-past helpdesk days I've been on the receiving end of more than one incandescent user ranting about my crappy network because 15 of them are trying to edit the "Sales Database" (an Excel doc with about 100 subsheets) at the same time across the network. This has happened even at companies who have purchased software for this very purpose to replace Excel that the users don't like and won't use, and no amount of explaining that the file created 15 years ago when all this was fields and there were 2 users isn't designed to scale to a multi-user environment ever gets the message across.
Excel is a prime example of the mantra "if the only tool you have is a hammer everything starts to look like a nail"
the file created 15 years ago when all this was fields and there were 2 users
And that's the point I was trying to get at earlier.
Because Excel is the go to start point for a simple data set it locks even large organisations in if they start with it, because they weren't a large organisation/department when they started.
In effect it's not Excel is a prime example of the mantra "if the only tool you have is a hammer everything starts to look like a nail" so much as that when you are using an awful lot more nails you might need a bigger hammer.
What's lacking is a simple ( Excel like) programme, that easily scales up without learning a new, complex and scary programme.
I suppose the dream would be something that still looked like Excel even when it had been turned into a proper database under the bonnet.
Some of the spreadsheets many of the PMO people I know have to use are insane and take forever to load because a network stored ridiculously sized spreadsheet has grown out of something once thrown together quickly to centrally store data from emails in. Some of them are also so complicated that no one knows how to fix them when they break because the person who made and hotfixed it has since moved on.
In such cases you have to wonder;
1) Was this actually the most sensible approach at the start, i.e. did they not think of creating something more appropriate or was this just one more example of that bane of our lives, short termism- get it done so that it works now and by the time it goes pear shaped they'll have moved on?
or
2) Was there a phase two point when they could have concluded that it was time to move to something more suitable while it was still practical to do so, but decided not to, because short termism again - bodge it so that it will last a bit longer - as above
. Or
3) Did no one ever even consider thinking about perhaps reviewing their current working practices in the light of increased complexity and size of the projects?
I've hit my head against the consequences of all three over the years. And in every case it was all about not making a (spending) decision in this year's budget. And I've met plenty of other similar situations too.
My (least) favourite example - the council decided to have all singing, all dancing copier printers everywhere, that could be networked to our PCs. But decided not to network ours, that year, to save a few bob.
Two years down the line when it was very clear that this was wasting useful amounts of ££ and time they, eventually, agreed that it could be done, and then they another eventually later agreed to actually allocate time to do it. Except it couldn't be done. It wouldn't connect to the network. And after yet another eventually they realised that the machine had been installed without a network card, to save a few more pennies, And by the time they'd worked this out there were no available cards, or contract supplier's will, or something. So it didn't get done within the lifetime of the contract.
No (apparently?) one knew who'd selected this or signed it off, of course.
Wish I could upvote you a hundred times. It is terrifying to think that a lot of businesses long term plans are just numbers and equations hacked away at by an accountant. One place I worked at as a summer job as an accounts clerk, they asked me to look at their cashflow spreadsheet because something wasn't adding up right; half the equations had been copied incorrectly and they had to revise a whole bunch of plans because even the numbers they thought were right were just plain wrong.
The thing with excel is that everyone knows how to use it. That belief in turn gives them confidence trying advanced excel features, when the same user would balk at learning to do the same with a database. It's not how difficult it actually is, it's how difficult the user perceives it to be.
The other thing is, businesses are unwilling to go to the expense of training so many users up, and giving them database licenses. It also takes time to get it exactly right, time which many businesses do not have.
Excel gets it done OK enough right now, which always trumps getting it perfect next month or next year, because many businesses can operate with a bit of long term inefficiency, but most businesses would even start up if they had to wait weeks or months to be operational
Talked a a Prof of maths. Guy owned a data science company, went were he wanted for interesting course to teach at universities around the world. Still said excel is good for quick and dirty and that even though they use tons of actual proper tools for mathematical modelling and data visualisation. They still tended to present it in excel a lot because it was a format everyone know when you are looking at numbers.
*nods sadly*
At [RedactedCo], we have some spreadsheets that our finance department uses that pulls in data from multiple sources, folds spindles, and mutilates it into a format that the analytical accountants can read through. Back in the early windows XP/7 days, we had to fit several of the computers that used these spreadsheets with more memory, because excel took all of it.
There was originally competition, but Micro$haft blatantly abused its monopoly position on the desktop and they became victims of bundling.
Difficult to dislodge it now so the world has one sub-par product, instead of a whole market segment dedicated to smb math/analytics.
Downvoted for the use of Micro$haft are we not all grown ups
We all hate Microsoft for what they have done over the years, but for the amount of hatred you must hold you must be well over the age of 30 and as such using such childish disses such as that are poor performance
Yes, Excel has limitations (as does SAP's product). But for 98.7654321098765% of users who can live with 15 digit precision, it's good enough. Not to mention the legacy code and user familiarity.
Perfect is the enemy of good and El Reg is filled with tales of failed development projects where "nice to have" requirements became "mandatory" in pursuit of perfection.
Besides, users don't want to be told "you're holding it wrong."
I think you mean there is no other application that comes close to a spreadsheet in terms of usability, versatility and usefulness.
Excel is one example of a spreadsheet, with quite a few good features, but also a whole load of limitations and bugs, which my fellow commetards have only been too eager to point out.
it can be made to do *almost anything* to varying degrees of success.
Many, many moons ago I was rental manager for a stage lighting company. The tech guy for the company liked Macs (we had even had a Lisa before that) and he bought a stonkin' expensive Mac IIx for the rental department "so we could keep track of our bookings." Problem was, there was NO rental tracking software for Mac at the time.
Enter Excel. I sat down with the printed manual (!) and figured out how to write macros that would subtract units from inventory over a selected time frame. Over a year, that evolved into a relatively mature and stable system with custom menu bar, dialog boxes, etc. that anyone in the department could use with minimal training. The company used it for years, even long after I left.
All done with Excel for Macintosh (1.0 & 1.1 IIRC), half a decade before Windows 95.
So yeah; I can see people being loyal to Excel. Sure, it makes everything numeric look like a nail; but you can accomplish a lot by hammering on things.
Every large corporate I've ever worked at basically runs on Excel.
Give people fancy ERP, analytics and BI tools and all these will ever get used for is to create data extracts to be manipulated in Excel.
Its always at best the second best tool for the job for anything, and users like its familiarity.
Yes, but the users aren't always doing this out of blind stupidity. A 'corporate' system always crystallises to a particular scope, and inevitably the outputs need to be viewed, or post-processed, or further modelled along with other non-corporate data. Not all the the data can be crystallised into the main system, as - in almost every case - these are the areas where the processes are being actively designed and developed to the needs of the decision makers. Sometimes this just never stops long enough for crystallisation to be feasible.
Excel enables a combination of roles - process design and decision making. This is almost always necessary when the process constantly evolves or never settles to a particular steady state. To separate these roles requires more resources and stability of the process, and comes of the expense of flexibility. And whilst an excel output lacks the control of a more formal system, these controls (validation, repeatability, ISO9001 etc.) can be built into the system by a competent enough end-user.
>Yes, but the users aren't always doing this out of blind stupidity.
Particularly, as once you've learnt how to do pivot tables, charts etc. in Excel, you can use that knowledge for any other package, whereas knowledge of SAP Analytics pivot tables is only applicable to SAP...
What's annoying is any spreadsheet of sufficient complexity is just as difficult to get right as spending the time to normalize your business logic into a database. The spreadsheet offers the deceptive simplicity of "changing this one thing here," but carries the risk of borking the rest of it. If your data is important, it's important to get somebody competent to build it for you.
Graphic designers have the same issue with Word and Powerpoint. It's easy for amateurs to put something together, which makes the amateurs feel empowered. But amateurs think clip art and a half-dozen different fonts makes something "creative and fun!"
Budgets are somewhat elastic items. Depending, as they do on a foundation of net cash availability, overbuilt with internal politics, wilful ignorance, actual incapability and favouritism. e.g. the £20,000 that has been set aside for software improvements has largely been used to provide the department head a new set of furnishings for the corporate meeting room, that is never used for meetings unless they want to impress someone with the contents of the fancy drinks cupboard.
*Personal experience- over the years my frontline educational support team were "gifted" a large number of nasty, useless and often inconvenient ( because blocked the entrance to our cramped premises - not for us the extensive office space that the "officers" skulked in) piles of junk. Because we needed computers, desks, etc. and the higher ups decided that their stuff wasn't good enough for their fancy corporate offices - but would be made good use of by us peasants who only did the actual assessing, educating, advising and training that gave them an excuse to have their much better paid jobs.
Those of us who were in the trenches in the late 80s and early 90s remember that Lotus 123 enjoyed a similar position of worshipful allegiance. We still remember the beancounters writing letters using one big row per line, because the answer was Lotus 123, no matter the question.
Excel may seem unassailable, but the history of computing suggests it's supremacy is just a matter of =TIME()
I've used Notes (casually, and largely as a tech doing installs and troubleshooting). I've used Groupwise back during the Netware 4.x / 5.0 days. I've used Outlook in all it's forms; We'll leave Outlook Express (later renamed to simple "Mail" for win10) out of this discussion.
They all that their good points, and at some level, they all suck various amounts of goat urine.
While I've never used the back-end of Notes and Groupwise, I imagine they are on the same level as Exchange.
I've yet to work with any application that does what those products do nearly as well and with that level of integration.
This post has been deleted by its author
"Excel may seem unassailable, but the history of computing suggests it's supremacy is just a matter of =TIME()"
There's a problem with this formula.
Not trying to type a formula?
When the first character is an equal (=) or minus (-) sign, Excel thinks it's a formula.
To get around this, type an apostrophe (') first:
Lotus 1-2-3 + Impress (or WYSIWYG) was incredible in 1992/3(?), just before Excel landed.
(For the youthful, prior to Excel 1-2-3 was character mode, and Impress or WYSIWYG allowed you to tab into a graphical environment familiar to spreadsheet uses today).
And Lotus Improv was out of this world at that time too....
I worked on a huge payroll system. We had a user who cost us a lot of time and money because he was phished with an Excel sheet that contained a macro virus. So security insisted on the introduction of code signing. Which was good. Within an hour an angry user appeared in my office screaming that I had "ruined" that weeks payroll run. I couldn't get much sense out of him so followed him to his desk where many excitable payroll people yelled at me, at lot.
It turned out that there was an issue with how the payroll system presented dates. To work around this the shouter-in-chief had designed his own Excel spreadsheet with a macro to change the date format and prepare a CSV file that was exported to the secure print system to print out the payslips. He'd then insisted that his minions use his spreadsheet. Only now it didn't work because his greyware had never been tested or approved by anyone and he sure as heck couldn't get his code signed.
I pointed out that it would take one of our DB people about, oh, fifteen seconds to change the database to display dates in the format they wanted and probably half a day maximum to get that through QA. But since it would be a trivial change that I could authorise the use of the code now and QA could follow on at their own pace. Not good enough for Mr Shouty. He wanted his code running *now* and wasn't prepared to accept that someone whose job it was to make changes could do a better job than him.
He took it to the CEO. I lost, in the worst way possible. I was told to revert to permitting unsigned macros to be run because that way Mr Shouty could continue to "Add value to the business" with his homebrew. I think they still use the cobbled together mess.
It’s been years since I looked into this but I remember that when the people who developed the OpenDocument standard went about reverse engineering the Excel format for interoperability they ran into quite a few issues that put me off using Excel formats for anything important.
The OpenDocument standards team received a lot of input from mathematics professors to prevent the mistakes that Microsoft had made. I believe they also had input from the European Spreadsheet Risk Interest Group (whose list of Spreadsheet horrors is a frightening read) to prevent common human error when working with spreadsheets.
There is a leap year bug carried over from Lotus 1-2-3. They screwed up the leap year maths and incorrectly considered 1900 to be a leap year. Not a problem if you never venture more than a century ago but if you have time series that go back to before 1900 your calculations will be off in Excel. (Excel wrongly assumes 1900 is a leap year)
I believe there was an issue with the order in which Excel performs certain operations (not entirely BODMAS?). If I recall correctly it would screw up mortgage calculations.
The challenge for standard developers is then to choose between perpetuating the error or lose roundtrip compatibility. For the leap year bug OpenDocument team chose not to perpetuate the error and MS now lists using the year 1900 as an incompatibility between .XLSX and .ODS formats (without stating that it's Excel that's at fault).
Anyone know more about this? The leap year bug is a WONTFIX, any others?
Related to the 1900 issue, is the borked epoch so if you format "0" as a date you get 00.01.1900 though the epoch is actually 1899-12-30T00:00 but Excel won't actually display anything before 00.01.1900 which makes it pretty unreliable for dates. Doesn't stop people using it for that, of course.
In fact the ability to mix both data and code is the biggest risk associated with any spreadsheet because you cannot guarantee values won't change the next time you look. Otherwise, bugs aside, it's about as good a report format as we've got, because at least it's typed and supports unicode.
I hate spreadsheets with a passion.
My business has an outside accountant, who for the most part we just use for verifying our records before submission and the occasional tax questions.
I do most of the accounting myself, I do a full reconciliation report etc.
Anyway, initially they asked me to use Excel to submit it all to them, and after cursing at it getting slow, and having my hair turn gray, getting confused trying to put it all in a spreadsheet, I decided enough was enough.
I found an open source locally run application that was built for accounting and works on any platform, it does everything I need perfectly, I love it.
The biggest complications with the excel format is trying to calculate sales from online, instore and whether that money is cash (most fun is when someone paid a bit cash, and a bit card for the same transaction) , that money is card etc and having it all play nicely and tally up.
Ofc, when cash is deposited, it is usually different to the total sales figure, due to change given and what-not. That's a PITA to calculate in excel, because you cannot do double entry accounting.
So, the new app, can calculate that automatically and it all works great, and even fixed a few common errors in previous submissions, due to confusion in excel.
It can even generate reports that are identical to what the account gives us after reviewing.
After telling all of this to the accountant, guess what they said?
Please submit your records in excel.
They would not even look at it.
At this point my head exploded and splattered all over the room, with bits of brain matter stuck to the walls. So, I just copied and pasted the generated reports into excel.
The only good of excel is that I can use Libreoffice to make the excel files for them.
We spent a fair amount of time with one particular project and placed it live. We found that the users (forensic accounting btw) didn't like summarised or rolled up dashboards and wanted access to the lower level ledger data in Excel export to run their magic.
Accountants love excel and no matter of tech will get them over this - including SAP or Oracle...
Seriously. You can use it to do enough small different things reasonably enough. Like you can use it to open a wine bottle, open a tinned can, or trim a branch.
Problem is when people try to use it for a bigger project, get to the level where it's the equivalent of where you should use a chainsaw to chop down a copse of trees.
Rather than go through the pain of speccing reports, messing around with headers footers etc. only to be told "we need the widget id on the item line and subtotals by tax period" two days before go live.
I just dump anything useful is a downloadable spreadsheet and let the users work out what they need in real time. Works a treat.
What does not work so well is using spreadsheets as primary input. I have done this rather than go through a massive data entry re-work of a semi manual system where most of the data was in spreadsheets we just read the spreadsheets. About 20 different variations in the format over a few thousand files, and, I will gladly de-sanguinate anyone who even thinks "merged cell".
It all went relatively smoothly - bet then they insisted on continuing to enter data via spreadsheets.
....if Excel does what people want was well as they want it done, let them use it?
Obviously, if you have 20 years' worth of data for (hundreds of?) thousands of accounts, it's not the tool of choice, but there are lots of other use cases. Lots.
When you have to make 100+ rungs of nearly identical ladder logic to input map things, excel kinda sorta works.
Text export the master rungs, notepad++ shenanigans to get it to CSV with easily changed areas, open in excel, copy and fill-paste to size, modify mappings, save back to CSV, strip out the commas with more shenanigans in n++, and pray to the spaghetti monster it imports back into Logix.
It's a horrible horrible bodge from years ago, that I thankfully documented for when the inevitable orders for units with different I/O requirements came in.
I have written plenty of vital tools in Excel. It is Business's hammer!
I do now use GDPR & Governance as a lever to kill Excel in the front line. If only most customers knew how much of their personal data was blasted throughout almost every organisation in Excel.
SharePoint for data input and SQL server for handling data with PowerBI as the display. I can then at least try to tell who has seen what.
Other tools are available.!
Spreadsheets are the original malware as they are very difficult to verify once the size or complexity is beyond dead simple. They are useful for certain tasks but because of they appear to require minimal training to be competent using they have an allure that makes them 'femme fatale' of the office suite. You can do more damage with a buggy spreadsheet than with any other office program. Worse it is often not easily noticed.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
