Do Google Spreadsheets mean the end of Java?

by Paul Browne

Or to be more accurate 'Do Google Spreadsheets mean the end of Java as we know it'?

Update: There is a reply to the (many) comments on this blogpost here: Java and those pesky Google API's

Think about this. Who pays your wages Mr Java-Developer-who-has-just-had-a-couple-of-years-at-the-top-of-the-pile? Clients, or if you're in a larger organisation , the business folks (i.e.'internal' clients). Do you think any of them care about Java? Do any of them know what Java is? All they want is to get things done, quickly , and with as few mistakes as possible.

These business people would be happy to run their organisations on Spreadsheets. Do you remember the cartoon where Dilbert convinced the pointy haired boss that he could fly the plane using Excel? There's more than a element of truth to this. I know of at least one US Fortune 100 company that (until recently) conducted most of it's operations on little more than Microsoft Office and duct-tape. It worked, not very well, but it worked.

Until now , the next line would be 'Excel (or any other type of Spreadsheet) is not secure / scalable / sharable / not web friendly'. That was until Google launched their Docs and Speadsheets. It's an online version of Office with some spreadsheet functionality. Play with it a bit and you'll see that there's plenty missing. But this being Google , I'm willing to put good money on

  • (a) new features rolled out (think steamroller) and

  • (b) These Spreadsheets being massivly scalable / secure / sharable.

This being Google, there is also an API (developer page here). It's got massive holes in it (e.g. you can't yet use it to create a new spreadsheet). But when Microsoft bring out their version of online spreadsheets (and they will) not only will they clone the Google API (to get market share), they'll need to go one further and introduce new features / remove the usage restrictions in order to compete.

So, secure, scalable, sharble online spreadsheets are here to stay. So lets take a look at Mr. (or Ms.) Pointy haired boss thinking about their new project:

  1. Hmm, I think we need to be able to gather which health plans our employees are enrolled in.

  2. OK, I'll throw together a spreadsheet to show people what I want

  3. Before I'll give to our friendly Java developer and let him 'do' a website from it.

    Soon I'll just share this on Google.

  4. Great , Loads of people are now using it, I'll just the (Ruby / PHP / Insert other language here) guy to add one or two extra features.

  5. Most Excellent. Why don't we spin this off as a Web 2 company and sell it to EBay??

There you have it, Massively scalable , Highly secure websites (see Google Authentication API), without needing to know anything about EJB, JMX , JBoss, JDBC or any of the hard won knowledge that us Enterprise Java Developers have built up over the last 7-8 years. I'm exaggerating, but not much.

What do you think? Is Enterprise Java dead, or is Web 2 just another boost and a slightly different way of doing things for us Java people?

Other Java Posts from Technology in Plain English

Some other notes:


2007-01-18 07:29:33
Way off base. Java in the enterprise isn't trying to webify spreadsheet functionality - it's presenting a relational database in a meaningful manner. Spreadsheets aren't relational (no matter how good you are with vlookup), and you can't control presentation like you can with a real application.

If you need something very simple and sharable, it's already been out there before Google Spreadsheets came along.

Even Google Base doesn't step on Java's toes (or more specifically, web application developers). It's just a different place to store the data. Now there are some MS-Accessish web applications that allow the users to build their own front ends to their databases, but in the end users can only do so much. They're limited by their incompetence, and moreover their time. Developers are good at this stuff because they've spent the time learning how to put things together and express themselves programatically. Users, on the other hand, hold their professional expertise in a different area (like Finance or Health Care). While some of them may have the requisite skills and insight, they've got other jobs to do.

Paul Browne
2007-01-18 07:43:32

Way of base? In a narrow sense yes , but looking at the bigger picture no: If I can capture information on the web more quickly with google spreadsheets, I can live without some of the advantages of Java.

If most of Java in the Enterprise is trying to shuffle information to and from the web and a database, what is Java doing that something like PHP or Ruby can't do faster? Why not replace the database with Google's version of Excel using the API to communicate with it.

Yes , there still will be a need for developers (users don't *want* to develop things), just not the sort of developers we have now!


2007-01-18 08:08:15
Why is it just the end of Java? What about .NET? Maybe it's the end of all programming languages. Why are Ruby and PHP still useful with the all powerful Google spreadsheets in existence? Hmmm... I wonder what language Google Spreadsheets is written in? Hope they didn't use Java, because it's the end of Java. Maybe Google Spreadsheets will implode.

Anyway... I don't agree with you at all. Google Spreadsheets is very cool and very useful, but I don't see how a cool web application all of sudden makes a programming language obsolete.

2007-01-18 08:13:19
Paul B,

Stepping back and looking at what my clients clamor for from Web 2.0 (web integration of legacy products, modernization of "good" mainframe applications, solving their business problem with manageable software), I don't think that Google can yet claim killer app status, especially with Google spreadsheets! So will Google spreadsheets kill Java, no. Not by a long shot for me. For my peers that program Java, they haven't been worried about Excel for the last 20 years, and they won't worry about Google Spreadsheets in the next 20. We worry more about .NET =P

On the spreadsheets-serves-business-needs side, yes Google now has joined the fray, amongst a myriad of spreadsheet apps. Has anyone out there besides me used Google's Spreadsheets? It is far underpowered for the power user. If Mom-N-Pop want to make a budget in Google Spreadsheets because their sewing shop can't afford Quicken, sure, they may use Google Spreadsheets (if tehy can get load it on teh Internets truck's bunch of tubes).

In short, I don't see this even competing with Excel.

Paul Browne
2007-01-18 08:19:25

You're correct - this affects .Net as much is it does Java.

The subtext is 'the end of java *as we know it*'. I think Java will still be around , but the way we use it will change - more emphasis on using other people's infrastructure, and less on building our own.

My bias in the post towards Ruby and PHP is that they are 'duct tape' languages (this is a compliment), suited to quickly (and maybe dirtily) tying together bits from other systems).


Paul Browne
2007-01-18 08:27:56

Take a look at the bigger picture. Yes Google spreadsheets is way underpowered, but do you think it is going to stay this way for long.

My key argument is that a full featured Google Spreadsheet API will offer the functionality that previously *had* to be delivered by Enterprise Java. This will fundamentally change the way we think about delivering systems.


2007-01-18 08:49:12
This is probably the silliest "Java" article I've ever read.

I'd love to see you tie a 40 year old mainframe/COBOL system into a multi-million row un-normalized Oracle database and an SAP inventory manager in a way that complies with privacy laws with an extensible, scalable security model across continents using a Google spreadsheet.

Can Google spreadsheets potentially provide a (yet another) quick and dirty way to jam out a functional web UI on top of a database? Yes, certainly. Is that cool? Sure.

ASP was allowing us to plop a web UI on top of an Access database or an Excel spreadsheet in 1995. Still Java carried on.

I'd say you're exaggerating a great deal. How hard is JDBC, really? Is it harder than learning a Google API?

You certainly don't have to "know" JMX for anything, unless monitoring and instrumentation is what you're going for (will Google spreadsheets even provide monitoring/instrumentation as an option? What if you find out a year after you launch your app that you need it, or would like it?). You also don't have to know EJB, especially with Spring.

Yes, you do have to learn things to use J2EE, I'll concede that point.

As far as paying for hosting, yes it is true that J2EE hosting is more expensive than PHP. But if your company is literally counting dollar bills (or Euro bills or whatever) and complaining about $50 a month versus $5 a month, then I hardly think you should consider that "enterprise" development. I'd also think very hard about working for that company and have serious doubts about how serious they really are in regards to their software.

Also, now many CIOs are going to feel comfortable placing their entire company on software that is in perpetual beta?

And finally, Google has a pretty long history of hooking their stuff into Java APIs. They've provided several Java APIs already, and I'm sure they'll do the same here. If anything, Google spreadsheets will become another tool Java developers can leverage in developing their applications. But I guess "Google spreadsheets mean you can get data in Java" doesn't make a very sensational headline.

2007-01-18 08:55:55
You are comparing an application to a language. Languages beget applications. Essentially, everything that an application does is possible because the underlying language made it possible. The application is therefore a subset of the language which created it, and can consequently never outdo it.
2007-01-18 09:09:02
>Take a look at the bigger picture

2007-01-18 09:09:08
Please could I get some of the good stuff you're smoking! Most organisations are trying to get AWAY from spreadsheets. The company I work at (a large finance company) looses MILLIONS every year from simple cut-'n-paste errors that occur in spreadsheet applications. Do you seriously recommend that developers put their efforts into learning Google's custom spreadsheets API over learning standards-driven (via JCP) or open-source API's like Swing or JSP.

You're as mad as a frikken hatter mate!

2007-01-18 09:13:57
Are we at Techcrunch here now? Long time since i read so much BS.
Google Spreadsheet is a long way from being anywhere near Excel feature-wise, and most of these features are heavily used. Good luck trying to reproduce this with a web app (and i like AJAX/Comet a lot) without requiring the user spending n times more effort in achieving results. Google Spreads can't even import a spreadsheet with more than 30 pages, so i still have to send around my excel sheets by mail. But enterprises rarely need this, i am sure.
And their confidence in Google got certainly a tremendous boost after they canceled their search API. I just see them crave to render themselves hostage to Google or the more reliable and customer-friendlier Microsoft. This op-ed is so ridiculous i am sure you will get a lot of reaction. What probably was your intention in the first place. Your calculation of ad revenue vs. cred spent could be flawed.
2007-01-18 09:32:52
All of our clients use Excel, but utlimately to tie in inventory, general ledger, customer orders, etc, you need a relational database and business logic. Spreadsheets don't offer those, no matter how sophisticated the api.

I'm of the opinion that Web 2.0 is highly overated, with regards to browser applications. Consider this: The latest Mozilla build is roughly the same size as the 1.6 JRE, and is of course far slower and less functional. Furthermore, you have to account for browser differences(STILL) when developing applications. Browsers are meant to... browse, not build applications. Browser applications are evolving into what Java and .NET already are. Web 2.0 is still just a network, which anyone can plug into.

You are right about one thing though, "All they want is to get things done, quickly , and with as few mistakes as possible."

Paul Browne - Technology in plain English
2007-01-18 09:49:55
JCN, Eoin,LG and AndrewN and Ihe

Thanks for your comments and to answer (part) of each in turn.

- JCN: You've correctly identified one of Java's remaining Niches as system integration. But is that the only Niche? Is Java destined to become a 'Legacy' system with only the cost of change preventing people from moving on. New systems will be built in fundamentally different ways. Some of these (hopefully) will be built using Java. But they won't be built using the J2EE API's that we have today.

- JCN: Cost of Java hosting - not a problem for the corporates , but for the wave of startup activity every penny counts. Once Java loses the 'bleeding edge' of new innovation, we may see a lot of the best people moving 'beyond Java'

- Eoin: Which is more important , the language , or the core API that various languages use? Microsoft Office (which at it's heart is an API that can be driven by various languages) has been far more valuble to Microsoft than Java has been to Sun.

- L.G.: The Register article is correct (having programmed Spreadsheets a lot in a previous life). But what if the major faults were remedied (as appears to be happening with online spreadsheets)? Should we not look again at some of the assumptions we've made as how we build things for clients?

- AndrewN: A similar point : Current spreadsheets are bad (for all the reasons you mention, and many more). But what if Google spreadsheets didn't have these problems? We don't try to build an OS any , we build on top the existing OS / API (e.g. Windows , Linux) and move on. Likewise for the new Web API's becoming available.

- Ihe: Google IS a long way from being an Office replacement. But already it is 'good enough' for a small number of users. I am confident that they will add more and more features over time. Any missing 'must have' features you as a developer will add. But this 'add the missing 10%' features is very different from 'lets build a complete financial application'.

Paul , Technology in Plain English

Paul Browne - Technology in plain English
2007-01-18 09:51:56

Read the subline. Google Spreadsheets mean the end of Java as we know it'.


2007-01-18 09:52:56
This is the most asinine "Java" article I have read in a long time. Spreadsheets replacing a programming language? Let us know when you come back from the magical mystery tour.
Robert Jones
2007-01-18 10:02:52
I know of at least one US Fortune 100 company that (until recently) conducted most of it's operations on little more than Microsoft Office and duct-tape. It worked, not very well, but it worked.

Let's see, you have one example, and it didn't work very well, and they don't use it any more. Spreadsheets are good for what they do, but they have their place. Once they get more complicated, and bigger, you need Microsoft Access or something.

Paul Browne - Technology in plain English
2007-01-18 10:03:15

Don't agree with the comment

to tie in inventory, general ledger, customer orders, etc, you need a relational database and business logic

It's a lot easier if you have a database ,but it's amazing how far you can get without one and paying people overtime to do the jobs that can be automated.


Robert Jones
2007-01-18 10:05:48
We've been flim-flammed by yet another character walking around with an "End of Java" posterboard pinned to their shirt. owel. [What's the best way to drum up "hits" to your web site? ]
Paul Browne - Technology in plain English
2007-01-18 10:06:23

You'd be surprised at how many Excel spreadsheets are out there. Typcially us folk in the IT Departments don't get to see them , because the tend to be 'user generated'.

Yes, spreadsheets are hidiously awful , but users are happy with them (and online spreadsheets will make them happier still)


2007-01-18 10:11:34
The end of Java? Google Spreadsheets was programmed using the Google Web Toolkit, which requires Java coding. A bit ironic for this article, don't you think? Take a look at this article:,289483,sid26_gci1218941,00.html
2007-01-18 10:27:30
most of it's operations -->
most of its operations
yuhong wang
2007-01-18 10:51:53
Have you done any real data analysis work? Please do some and then say this.
2007-01-18 11:43:11
Dude, way off base. Spreadsheets are 1980's technology -- if they are "good enough" to suit any enterprise use, then why didn't we just stop there? Every tool has its merits, it's a good technologist who picks the right tool for the right job. This is just as dumb as saying that .NET is always better than Java or vice versa. Just because my boss "gets" spreadsheets doesn't mean that 4GL languages don't have a valid place in our ecosystem.
Paul Browne
2007-01-18 13:19:45
Again , thanks for the comments. To answer them (in some sort of order).

-Robert. In Terms of hits, I don't see the figures. O'Reilly kindly allow me to blog my sincere opinions (which you may or may not agree with!)

- Jorge 'The end of Java as we know it'. Will we see a small, hard core of Java programmers a Google writing a Java Kernel that everybody else uses via some sort of scripting language? (think Linux Kernel here - a lot of people use it, but only a small number of dedicated people actually write it).

- yuhong wang. We're talking early days here, so any figures that are available (e.g. growth in use of Ruby against the Google API's) could be extrapolated to say that everybody in the planet will be coding in that language by 2009 - an obvious misuse of the figures.

- Nemesis. Spreadsheets may be 1980's technology, but there is a vast difference between Lotus 1-2-3 on the original IBM PC , and on online sharable repository that happens to use spreadsheets as its user interface. A key point is that we (as in technologists) may never be asked our opinion on .Net / 4GL or the merits of any other technical solution as the 'updated' spreadsheet solution is now good enough.

2007-01-18 13:28:56
Once again a nothing story with no substance.
Computers is all about human interaction stupid!
Willem van Es
2007-01-18 13:46:56
Thank you for wasting my time.
2007-01-18 14:24:37
The ultimate answer to the question posed in the article is "No".

The "as we know it" fine print is irrelevant. We've already seen several ends to "Java, as we know it". Java is growing and changing and continuing to evolve. OSS Java is the end of Java as we know it. Swing was the end of Java as we knew it.

2007-01-18 14:37:36
This article is assenine, there may be an argument to be made about Java but this certainly isn't it. I don't care how many spreadsheets are out there. No one wants the user experience of a spreadhseet. Spreadsheets aren't going to provide the auditing history necessary, reports, handholding of an application etc. Quick and dirty solutions often fail, that's why people take the time to design and build well thought out custom applications. And they will continue to do so without any changes to java regardless of the success or failure of Google Spreadsheets... Regardless, I'm replying to such an assenine idea I hardly know where to begin. Was this article really thought through? You certainly can't believe this Paul, can you? Are you trying to curry favor with Google for a job offer or something, that is really the only thing I can think of. Anyone who claims Google Spreadsheets is so powerful it could display modern programming languages as we know them, you must think they're going to send you a Google Switch or something for your troubles...
2007-01-18 14:42:55
Your argument may have some merit if Google spreadsheets were enterprise ready. Try uploading a spreadsheet that is bigger than 500K. Try uploading a spreadsheet with multiple images embedded in multiple sheets. No go. So as it stands now, Google spreadsheets has a long way to go to replace excel, much less replace java.
2007-01-18 14:51:14
First of all, as many others have commented, this would apply to any language, not just Java. Secondly, we are speaking of the presentation of data. If you are planning on storing all of the data for a company in spreadsheets that is fine, but good luck running any unforeseen queries (that is what a relational database is for). However, software is about much more than modeling and presenting data. It is also about manipulating that data. For example, a system that allows ordering items (amazon for example). I don't think you will ever be able to replace amazon and all of it's back end services with excel or some other spreadsheet.
2007-01-18 14:51:41
This is hilarious.
2007-01-18 14:57:43
The hardpoint of google spreadsheet, as well as many other web2 applications, is not in the feature richness, but in theyr...web nature...
I can access my spreadsheet (or document, to-do, calendar, name it yourself) from everywhere in the world. I can enter a net-cafè and continue writing my novel, or ask a friend to let me check my calendar from his computer. THIS is the power of web2: to be accessed from everywhere.
James D
2007-01-18 16:24:45
While you may be correct in assuming that Google spreadsheets will find a niche for some very specific cases which are implemented today using Java, to say that it spells the end for Java is simply ludicrous. You made some very big leaps of faith in this article with very little facts to back it up. I think you'll be busy explaining yourself to all the posters here..
2007-01-18 16:35:17
Wow, what a load of garbage!!!!

First of all, the biggest problem plaguing my Fortune 50 company is we tied too much with MS Excel. Yes Excel is cool and user-created spreadsheets save time and are flexible. However we've sunk so much intellectual property and time into Excel , we're screwed and can't back out to use something like OpenOffice. Now when our Execs try to bargain with Microsoft for a multi-million dollar license agreement to use the newest Office suite company-wide, they've uped the licensing price 10-fold.

Propietary Vendor lock-in is suicide and this is no better example. Even if Google makes the system open we're susposed to move our company IP and data into their systems? Sure, right.

Google has quietly become the new Microsofrt (big brother). They have some great products, but just because they come out with something new and innovative doesn't mean it's a replacement for tried and true proven technology.

Let me grab MY Magic 8-Ball and start making predictions. How about Google buys Oracle and Microsoft in the same day. Oh yeah, Google has a partnership with NASA - how about if they open a Google office on the moon?

JCN : Rock on.

2007-01-18 16:36:56
Put the crack pipe down, please!
Yeah Right
2007-01-18 17:17:59
I don't even like Java and I think this is the stupidest thing I've ever read. But what can you expect from a dude in a tie who claims to know something about software development. Ouch. Sorry, that was unreasonably harsh. But this is still dumb. ;-)
Sviergn Jiernsen
2007-01-18 18:16:32
Dear Lord! Imagine if medicine, trading in people's health and lives, were done the way several of the people describing the business application development process here seem to be suggesting. "We have a new sexy-looking bandaid that will make the problem the patient is having seem to disappear, quickly and easily and cheaply, but in the most superficially suboptimal way imaginable, making it impossible to fix problems down the road. Whaddaya say? Shall we go with it?"

The sad thing is that I fear for our lives that medicine has already succumbed to becoming this business-like, which is not a good characteristic to have, for medicine or for business.

Tony Williams
2007-01-18 20:53:56

You have a strong point and half your readers will deny it till their death. The truth is that *development* as we know it is already disappearing.

We are an incredibly young industry and continue to learn and innovate in the ways we deliver what users want. More and more the solution that delivers three quarters of a users needs in a fraction of the time and cost will win the user. Online apps like Google Spreadsheets and Dabble are part of this.

Case in point :- A large organisation I worked for recently decided that they would shut down use of FileMaker Pro. The current development team (which uses Java and either Oracle or Postgres) quickly discovered they had enough work for two years just converting the databases central IT knew about. A user survey discovered they were running three times as many applications as they thought they were as users had "thrown together something quick till the development guys can get around to us" and kept that solution for many years.

I've also seen users keeping several thousand contact records in an Excel spreadsheet or several hundred record mailing lists in Word. All because it's what the user knows and it does the job.

Us developers and geeks think users want all the power and functions of the big time stuff. Truth is users just want to do the job.

# Tony

Paul Browne - Technology in plain English
2007-01-18 22:58:06

Thanks for the positive comment :-)

Given the nature of the reaction to the piece (3 positive reactions out of 34), I'll accept that the way I'm trying to say things is obsuring the message.

The message is broadly 'hey guys, there is new stuff going on out there, outside of Java. It's not guaranteed that us Java guys are going to come out on top this time around.'


2007-01-19 00:07:38
No Paul, this article is full of crap, I don't even know what's your point.

No matter how advanced Google Spreadsheet becomes, it won't be the end of Java as we know it. Never.

2007-01-19 01:55:33
Who would be so insane to give really important data
to some company especially if it is subject of a government
obsessed with terrorism paranoia :)

2007-01-19 02:08:53
Google Spreadsheets is written in Java!
2007-01-19 02:14:45
Actually the title of the article was "Do Google Spreadsheets mean the end of Java?" I've seen people talking about Windows vs Linux, Java vs .Net etc etc but comparing Google Spreadsheet with Java is simply stupid and the Answer is NOOOOOOOOOOOOOO, it is not the end of Java (as we know it).
IT Expert
2007-01-19 02:19:11
I think its google maps that Java people should be worried about...
Paul Browne
2007-01-19 02:36:28

I've seen people talking about Windows vs Linux, Java vs .Net etc etc but comparing Google Spreadsheet with Java is simply stupid and the Answer is NOOOOOOOOOOOOOO

I'm sure the people in IBM said something about the PC. 'Its such a dinky little toy, how can it hurt us'. Just because something is radically different doesn't mean that it can't radically change things.

Google Spreadsheets and Java both exist to solve user problems. If I can solve the problem a better / easier / quicker way than Java , shouldn't I use that?


2007-01-19 02:42:09
What a load of b***sh**
2007-01-19 03:35:41

You're not a popular man.

If you are trying to stir up controversy, then you've done it.

I understand the point you are trying to make. The real key is that would any CIO in their right mind hand over their entire IT kingdom to google? What is the Business Continuity plan when Google Spreadsheets goes down one day? What if it goes offline for a few hours? Sure you could have a desktop client / way of persisting...but what happens if you're a global fortune 500 company and it can't wait? It's not like you can call Google and say 'can you please get everything back up, we are completely and totally reliant on you, so work faster.

Some smaller org's can possibly do what you are suggesting...but it's absurd to think that everyone could.

2007-01-19 04:50:18
What are you thinking? The spreadshit(ops, should be spreadsheet) was created long time before java came out, why jee get its way into the company? Didn't you write a blog said that the java/jee is "Excel terminater"?
Frankly speaking, I don't think a compnay can build on top of spreadsheet(purely spreadshit office?), can you show me one?
2007-01-19 05:09:17
I think you are trying to start a war of java and spreadsheet, but I think you are terreblly wrong, the title should be "Does google spreadsheets means the end of Oracle?".
Back when I was in another company, there was a spreadshit guy, the guy spent most of his time on developing "spreadsheet", he use spreadsheet to log everything (code changes, reasons...) and then put whole bunch of spreadsheets into CVS, and when people check out, 90% junk are spreadsheets, and people don't what those spreadsheets are. and of cause the CVS can't be merge (toooo big).

THat's the most stupid thing that I had to live with, I can't believe people like you can think of spreadsheet replace java.

Oh, btw, JEE is not only means web, web is just part of it, think of schedule, SOA, WEB Service.

I'm wandering that if Google paid you to do this or you are totally dumber?

Paul Browne
2007-01-19 05:24:05
Once again, thanks for the comments (both good and bad!)

- javaguy44 : Controversy is (sadly) getting in the way of the point I'm making. Your point about single supplier is good. But what about when Microsoft clone the Google API (as will sooner or later). Would you be happier to consider this approach then?

- coredump : Newer doesn't always mean better. In some ways the Web is just a polished version of mainframe technology (big central computers, dumbish clients). Google Spreadsheets, Web 2.0 and the API's may just be a polished version of Excel, but if they fix the problems of 'old style' spreadsheets, should we not look at them to solve problems?