InfoWorld: Thin vs. Fat: Google’s plan to kill Microsoft Office
Case in point: Microsoft Office. Most people think of the big three — Word, Excel, and PowerPoint — as merely an integrated suite of stand-alone applications, albeit a wildly popular one. Take a closer look, however, and you see that Office is much, much more. Thanks to the inclusion of some robust integration APIs (Visual Basic for Applications, OLE automation, and various add-in interfaces), Office is a commercial development target in its own right. In fact, one of the easiest ways to break into the Windows development marketplace is by targeting Microsoft Office. Make it do something new or better and the world will beat a path to your door.
If you can make a list, you can make a blist. If a blog is a web log, well, a blist is a web list. A blist is a list with some structure.
Didn’t we have the thin vs. fat debate in the early ’80s when we sat at dumb terminals allowing a ne’er seen Cray super computer compile our “Hello World” Pascal assignments? Didn’t fat win? Is this just where we are in the cycle?
I’d like to say cycle. I’d like to say that, while thin’s resurgence is obvious now, it will peak at some point and we will gradually migrate back to something that resembles fat. The problem I have is predicting what it is that will pull us back. What is it about fat that will bring us back to it and why isn’t it preventing the move to thin now? I can accept the argument that I’m simply framing this argument wrong. Maybe some things will stick with thin and others will never make it there.
Let’s talk spreadsheets. I’m pretty confident that no one reading this blog is ready to ditch Excel for an online spreadsheet. What are the issues? What are the must-have features in a spreadsheet?
The number one problem for me is being able to use it offline. There are very few instances where I’m not able to connect, but airports are some of those instances. I’m not paying $8/day to connect to the internet on a 1.5 hour layover. I’ve got to be able to get to my documents and executables in every situation – even if that means I have to plan which documents I’ll need.
For must have features, data validation, decent chart rendering, some sort of scripting, pivot tables, and array formulas come to mind. That’s beyond the basics of what makes a spreadsheet a spreadsheet. I’m sure there are others I’m not thinking of.
I don’t need collaboration, but I do need a templating system. I need to be able to have users create new spreadsheets from an existing template. Do any online spreadsheets have that now?
Other issues are privacy, security, accessibility to name a few.
For example, I work in a regulated industry -Biotech (Pharmaceuticals, Medical Devices, Blood Banks, etc.) and given the strictures placed on companies by the FDA, very few companies would consider keeping electronic records at a third party site.
TechCrunch’s Blist article
What you say Dick is true. But its not about what users need its about how much money may be extracted from various groups of users. ( Like splitting XP Home and XP Pro or losing Front Page when switching to Office 2003, etc, etc, etc). Most ‘users’ may not need a lot of bells and whistles but those that do may be compelled to pay more for being in the ‘minority’ – a la divide and conquer
I would add privacy to the list of must haves. A lot of companies are concerned about privacy, and I’m seeing more and more emails with boilerplate confidentiality statements beneath the signatures. Even if Google does everything it can to ensure users that their information is confidential, there will likely be a distrust of putting company data on some other company’s servers.
I also think there’s a tendency to think that the world is just the developed nations where high-speed internet usage is fast becoming standard.
Here in Russia, and I suspect lots of other countries (I saw a copy of Word(bootleg or not) in North Korea recently) the use of Office applications are widespread, however outside major cities, the Internet is unreliable plus, as Nicholas said above, it comes with a whole host of confidentiality issues.
I’m working on a laptop about the size of a dvd case here it cost 400 USD (Asus Eee). It runs a full fat client (open Office as it happens). It so easy to carry all our stuff round now (data and apps), I think the thin client argument is getting weaker and weaker. Maybe miniaturisation or convergence will bring fat back in fashion?
Many of us don’t actually need to collaborate as much as these cloud fans seem to think.
I think companies can get their Google apps on a google appliance inside their firewall so that resolves some privacy/security issues. Still leaves them with big feature compromises though.
Simon has hit on a point that rings true to me.
“Many of us don’t actually need to collaborate as much as these cloud fans seem to think”
I think the spreadsheet collaboration needs of business is both exaggerated and confused with other needs. First, I don’t think the word collaboration has come up organically at any client I’ve ever served. This is a fantasy that I just don’t see occurring in business.
The closest thing I’ve seen to what could be considered collaboration is the old data entry template which is fed by lots of people and funneled to one central document or repository. In those situation, I move *AWAY* from spreadsheets to a more server based utility.
I also think that Excel’s charting functionality is way underestimated by those who support the thin client model. I have yet to encounter a web-based charting product as robust Excel’s. Even SQL Reporting Sevices and Dundas charting falls short.
Why will we go back to fat client in the future? The same reason no one embraced the paperless office. We humans want to feel like we’re in control.
I don’t use any google apps for work, but I do use them just about every day for household organization and family business. My wife and I use the collaboration all the time. Even to the point where she is at home working on a document and I’m editing it simultaneously at work on my lunch hour. I’ve even used collaboration in the context of organizing an activity for one of our social groups.
Maybe there is not a lot of potential for google apps in the corporate space, but I find them very useful for my personal interests.
First off distinguish between the mainframes of old, terminal servers of today, and web-based applications of today and maybe tomorrow.
Mainframes lost out to PCs because mainframes couldn’t provide true interactivity. The ability to scroll through a logical panel larger than the physical monitor on a high-end terminal was nowhere near the functionality of Visicalc. Also, the only programming tools available outside MIS/DP departments were (my own experience) CLISTs, REXX scripts and VS BASIC. Power users couldn’t make mainframes do what they could make PCs do, and MIS/DP folks weren’t interested in multiplying their workloads by taking on department-level application development.
As for connectivity, many of the applications I use on a daily basis are hosted on terminal servers I access using a Citrix client. It’s not perfect, but I could access those applications (some of which are Excel workbooks) from any computer with an internet connection and Citrix client software. I already can’t do most of my job without an internet connection.
Scriptability is the main missing piece. Array formulas already exist in Google Spreadsheets and EditGrid. Templates already exist in EditGrid.
Some more of the driving factors in cloud/thin/fat are:
Telecomm bandwidth, price and availability
Hardware Prices
The price of software
The first two are driven ever faster and cheaper by Moores Law.
But MSofts dominant position tends to push the price of SW in the opposite direction.
We have probably already reached the crossover point.
This is an unstable situation that cannot last for long without Cloud and/or FOSS changing it.
One benefit to online applications (especially free ones) is that you can quickly setup an example file to be shared over the web.
If you want to show someone a data layout / formula or ask for help on this site (or any of the other great spreadsheet help resources) then just add a link to your post.
There are still a lot of missing features, but the apps are already a lot better than they were just 6 months ago.
I think that fat clients will be the mainstay for many years to come, but 10 or 20 years down the line, who knows…
Put it a different way, the major web spreadsheets are already LOTS better than the spreadsheet in Microsoft Works. This doesn’t really matter to Excel users/developers, but it does provide a strong indication that many, perhaps most, people who pay for Works are irrational, economically speaking.
As for privacy, security and accessibility, certainly accessibility in terms of offline use favors fat client software and locally stored files. But in terms of multiple possible users, web-based spreadsheets on thin clients are clear winners, no?
Privacy is only as good as user account passwords for the web-based services. For fat clients, privacy is only as good as the difficulty of stealing disks/computers. How many of you know someone who’s had a laptop stolen?
Security of fat clients is more a question of controls on users. If a disaffected employee could write client databases to removable media, how secure are companies’ files? Given the regular news reports of lost or stolen customer/taxpayer records, how less secure would society be if such data could be accessed from any web-connected computer over encrypted channels and only after authenticating nontrivial user credentials?
On-site data storage is no guarantee of security or privacy, and accessibility means many different, sometimes incompatible things.
Again, the comparison of current era fat vs thin clients to dumb mainframe terminals vs PCs of 20 years ago is just plain wrong. Besides, the uncontrolled, isolated PCs of the late 1980s and early 1990s didn’t win. Controlled PCs connected to relatively dumb servers did. Yes, PCs can still work even without their network connections, but when disconnected they do a lot less of all they’re used for now than they did even 10 years ago. But the key change over the last decade is that servers have become much smarter, and server-based software will continue to improve at a faster rate than fat client software going forward.
The only thing that might alter this is wholesale improvement in fat client software, meaning smaller, faster, more accurate, more tailored features. You know, software that converts 100-fold or so improvement in hardware processing speed into 10-fold improvement in user-perceived software processing speed. That would require a rather thoroughgoing revolution in software development, not least within Microsoft itself. That is, a decision NOT to add features but to improve the performance of existing features. Like that’ll happen.
Dick said “The number one problem for me is being able to use it offline.”
Consider long term. Not only supporting offline is a minor technical challenge, but you can see it that it’s obvious that web browsers, and mobile-based web browsers to a larger extent, are now carrying runtime execution environments, such as compiled Javascript, or Silverlight’s .NET. Both of these can support a debug/edit window which will be the exact replica of today’s VBA IDE.
So what’s your point against online spreadsheets exactly?
If you don’t think the fat Excel is dead, keep dreaming. This era is dead.
Dick said “The number one problem for me is being able to use it offline.”
Consider long term. Not only supporting offline is a minor technical challenge, but you can see that it’s obvious that web browsers, and mobile-based web browsers to a larger extent, are now carrying runtime execution environments, such as compiled Javascript, or Silverlight’s .NET. Both of these can support a debug/edit window which will be the exact replica of today’s VBA IDE.
So what’s your point against online spreadsheets exactly?
The era of fat MS Excel is dead.
I doubt fat Excel will ever go away. I’ve worked in far too many corporate environments where the idea of putting your sensitive spreadsheets (or should I say business logic) online with a 3rd party will never happen. And, believe me, all data is considered sensitive to these guys. Another reason why this will never happen is that there are far too many sophisticated Excel spreadsheets that have hooks into enterprise software and databases. Even if there was a way to ODBC into your corporate database via Google Apps, would you really trust putting direct access into your servers onto a hosted app? And beyond that, even less sophisticated spreadsheets often have hooks into that data through add-ins. I’m sure at this point that every mainstream accounting package has an add-in to allow end users to query data right into Excel. Good luck telling the accounting staff that they need to run their queries manually and then copy-and-paste into an online spreadsheet. And I doubt Oracle or Essbase or whoever is going to put out a plugin for Google Apps and Zoho and every other bit player in that market.
The only way I could possibly see this flying is if the online spreadsheet was sold as a product, not a service, and hosted directly on corporate servers. This kind of sounds like the direction MS is headed, so I would expect Excel would be a part of that in some form anyway.
Zach said “This kind of sounds like the direction MS is headed, so I would expect Excel would be a part of that in some form anyway.”
Yeah, sure, Microsoft is doing so many things at the same time. They have a gazillion incompatible APIs for data access, and keep coming with new ones every year.
As for storing stuff on the server, do you know that with the new so-called XML file formats, Microsoft is trying to get you to store a cache of your data inside files (they call this : Custom XML parts) ? With that in hands, do you think Microsoft is right headed or not, I mean from a data confidentiality stand point?
Offline support will be added to the Google Apps via Google Gears (gears.google.com) — It’s one of the most requested features). With offline support, users can choose whether to host their data online or not (allaying privacy concerns).
It’s not strictly fat vs thin anymore. The advent of Ajax and similar web/javascript technologies added a lot of “fat” to thin and Microsoft has been adding “thin” (online components) to fat. The web allows easy trial, distribution, collaboration and instant updates of software — it also typically carries a low monthly price tag or no price tag at all.
Yes, it’s true that you can set a reference to half a dozen different APIs and access external data half a dozen different ways with half a dozen different syntaxes. And this can be multiplied by the different versions of the APIs and again by the different versions of Excel. It is a PITA, and it makes it all the more difficult to share that workbook with other users even on similar configurations. But the point is, at least I can do that. There is just no equivalent in the online apps to do so.
To put it another way, there is no online equivalent of Crystal Reports. The closest we get are weak web query tools that are built into some enterprise apps.
On a different note, the one advantage I had hoped for in the thin app that has not materialized is speed. It’d be pretty sweet if I could load up a calc-intensive sheet onto the online app and let all the processing power the cloud could muster shoot out an answer as quick as a page refresh. Instead, doing it online is about a s speedy as doing it on a 512MB Vista box.
On the plus side, at least the thin guys are constantly improving. All fatty Excel has given me over the past 7 years is 2 steps forward 10 steps back.
Zach said “t’d be pretty sweet if I could load up a calc-intensive sheet onto the online app and let all the processing power the cloud could muster shoot out an answer as quick as a page refresh. Instead, doing it online is about a s speedy as doing it on a 512MB Vista box.”
That’s true right now. I think it will be addressed in the future though.