I'm tired of waiting for Google spreadsheets, so I'm building a fast alternative | Hacker News Reader

I'm tired of waiting for Google spreadsheets, so I'm building a fast alternative

hrishimittal.com

46 pointsrevorad57 comments

As an exercise, it's an interesting project. As a product, "better than Google docs" is not getting anyone anything - [Microsoft hatred aside] that space is where Excel lives. And as always, Microsoft is eliminating any of Google's potential advantages systematically.

After nearly 40 years of commercial spreadsheets, the empirical evidence is that spreadsheets are hard, otherwise there would be more of them. And the reason Excel dominates is because they get so many of the non-trivial details right.

Example: consider how often inaccuracy of JavaScript floating point operations makes a blog post appearance on the web [I recall several even on HN] versus the rarity of similar observations regarding Excel calculations. Then consider how many complicated floating point ops are done via Excel functions every minute of every day by non-computer scientists and non-programmers.

A couple of years ago I came to the conclusion that Excel, not Windows, is the killer-app for Microsoft, and that Microsoft knows it and invests in it accordingly. The person who decides on office suites is the manager who uses spreadsheets not the people using the word processor [in which case we would talk about WordPerfect not MS Word as the standard].

Sounds like someone who should be involved in the beta test to me.

I do not know what you mean by "Microsoft knows it and invests in it accordingly". For me, the big quality of Excel is that it has almost not changed in 20 years. If Microsoft changes anything in excel, this will hurt millions of users.

> I do not know what you mean by "Microsoft knows it and invests in it accordingly". For me, the big quality of Excel is that it has almost not changed in 20 years. If Microsoft changes anything in excel, this will hurt millions of users.

Microsoft has changed lots of things in Excel. The pivot tables and graphs were rewritten in 2007. The statistics functions were rewritten in 2010. These are pretty fundamental features.

The key is that they made these changes by versioning. v1 Pivot Tables are not automatically converted to v2 Pivot Tables. The new statistics functions have new names -- the old statistics functions are still around. This ensures bug-for-bug compatibility with old spreadsheets.

The UI of Google spreadsheet works much better than Excel (not to mention OpenOffice). No such nonsense as, for example, you can only paste something you copied once.

The reason Google Spreadsheet hasn't taken off is lack of investment on Google's part, which is needed because from what I understand from some Google's devs comments, it is a right mess under the bonnet.

We can only guess why Google haven't invested in it, maybe they don't believe people are willing to share their private or corporate data with Google's, or they think the potential for disaster is too high.

But I believe the potential is there

> The UI of Google spreadsheet works much better than Excel (not to mention OpenOffice). No such nonsense as, for example, you can only paste something you copied once.

I'm not seeing this.

You can indeed paste from the clipboard multiple times in Excel. I just tried this, and it works fine in both Excel 2013 and the Excel web app.

Yes but you can't paste anymore if you do anything else between copying and pasting. On 2010 at least.

The reason I've been thinking about spreadsheets is because of the potential.

I suspect the reason for Google's lack of investment is the degree of difficulty which it takes to develop a spreadsheet that can compete with Excel on something other than price. The world is littered with trivial attempts at spreadsheets, from Emacs org-mode tables and calc-mode to Common-Lisp's Cells project to Scheme-in-a-grid.

As I skim and wade and slog about Knuth's Semi-Numerical Algorithms, I get the feeling that any useful spreadsheet is probably going to have a bit of the "Electric components by Lucas" about it in order to bash a numerical implementation sufficiently to meet the reasonable expectations of real world non-programmer end-users. You won't get far with the CEO of a Fortune 500 corporation by saying, "Sometimes addition and multiplication are non-associative, so be sure to avoid cumulative rounding errors."

Pasting: See Office Clipboard. http://office.microsoft.com/en-us/word-help/copy-and-paste-m...

We're not expecting to steal any die-hard fans away from Excel. We couldn't possibly pry Excel from their cold dead hands :)

From the signups so far, it appears there are plenty of people looking for an alternative, so we'll focus on satisfying them.

I think that the empirical evidence shows that spreadsheets are hard to learn, and that unless people can take their hard-won knowledge to a new product, they aren't going to switch. Or that there are a lot of other types of software where you can make money but don't have to compete with the MS Office monopoly. Or, maybe, the empirical evidence hasn't shown anything conclusively about spreadsheets that it hasn't shown about word processing, presentations, web browsers, or photo editing: if you're the legacy brand of a complex industry standard, you have to be really awful not to own the market just based on your own inertia and deep pockets.

>the rarity of similar observations regarding Excel calculations.

There is no rarity of these: https://en.wikipedia.org/wiki/Numeric_precision_in_Microsoft...

>Excel, not Windows, is the killer-app for Microsoft

Agreed:)

For more than a decade, Lotus123 was probably more dominate in the spreadsheet business than Excel is today.

The thing about Excel is it's good enough to compete with free, e.g. Google Docs and OpenOffice/OfficeLibre. That's perhaps why Microsoft Excel buried less expensive competitors with approximately equivalent functionality for a typical user, not only Lotus123 but also Quattro Pro.

We use excel exclusively at our business and have tried Google Docs, but they have been so slow and break so often that we basically took the model we have been using online back into excel!

Google Docs scripts are cool but far too often they break.

In the interest of speed, what are you going to keep vs. trash out of the excel functions kit? I bet you could get an incredible tool by keeping top 30-40% of functionality, because lets be honest, anything beyond that an back to Excel it is...

We're going to start with the basic Math functions, sort and filter.

Another key feature will be a robust API to allow people to build on top of it. More on that in another post soon.

I remember quite precisely having collaborately worked on a single Excel document stored on Skydrive. I could instantly see the modifications my coworkers made on the document.

Maybe it works if both users are using the desktop version of Excel. But I am sure it doesn't work if one is using the web version and the other is using the desktop. That screenshot is from last week.

If anyone can confirm it's not the case, I'm happy to update the post.

So how are you solving the problem of connecting desktop app to web app? Are you writing an Excel add-in to propagate and receive changes? Or making changes to OpenOffice?

P.S. You complained about the selected-cell animation in the web app. This animation is not limited to the web app -- exists in Excel 2013 as well. It's just an animation -- there is no functional effect. If you don't want to get slowed down, ignore the animation and start typing.

It's just like unminimizing a window in the Windows taskbar. The animation is just eye candy -- the window is actually fully-active as soon as you click on the taskbar button.

Are you writing an Excel add-in to propagate and receive changes? Or making changes to OpenOffice?

I'm not fixing those apps. I'm building a new one without that issue.

This animation is not limited to the web app -- exists in Excel 2013 as well.

Yes, but it affects the web app a lot more. You're right that it has no function, so why slow down the user for eye candy?

You're writing a new web app and a new desktop spreadsheet? Very ambitious. I'll check back in ten years.

As for animations, users like animations. Only a small fraction of users know the program well enough to outpace the animations. And they can simply ignore the animations, like I do.

Both desktop versions indeed.

Office 365 doesn't allow two people to work on the same document at the same time? That's unbelievable. Isn't that 70% of the point of online document editing?

It does but only from the web app, not across web, desktop and mobile clients, as shown in the screenshot in the post.

Seems to me that I've had multiple people editing an Excel spreadsheet that was shared in a Dropbox folder, and it worked. Maybe we just got lucky, also it was a pretty simple spreadsheet, basically a task list.

Yup, Office, including Excel, allows up to about 5 users to collaboratively edit a document hosted on a local drive or SMB share. This has worked this way going way back -- as far as Office 1997 I believe.

I think it uses a lockfile to do this. It works, but it doesn't scale up well. IIRC, after about 3 people, the effectiveness drops substantially.

The restrictions are only for Excel. The other apps support many users.

Google spreadsheets have been neglected for a long time.

The sad part is that if this new project succeeded, it could catch google's attention and they will be able to launch a better version in a very short time.

They have too much power to strangle/buy any innovation. Just to let them die.

This is why I don't like google anymore.

This project is not dead. I cannot share exactly what's happening, but some cool new stuff is coming for spreadsheets.

Please make it something that helps me spot errors, and enforce better structure on the huge 2-d field of rectangles that has been in place since the 80s.

Honestly, I wish we'd see more spreadsheet-like apps move away from the "massive unstructured ocean of rectangles" and give us a little structure - something like Lotus Improv:

http://en.wikipedia.org/wiki/Lotus_Improv

mysimplegrid is a piece of attempt. Still quite linear and simplistic for now though. (http://mysimplegrid.com/)

Not meaning to hijack the post, but I am currently working on a solution for error spotting in Excel- we really feel that this is one of the other huge problems with spreadsheets. Collaboration intrinsically leads to "lost in translation" errors (see JP Morgan's recent blunder) and these are difficult to identify, especially when you are not the creator of the spreadsheet. We are at http://www.slateforexcel.com

I like it.

Yeah, google drive spreadsheets are only good for sharing fairly static small sheets. I just wrote an intro data science book that uses spreadsheets to demonstrate certain algorithms (simplex, kmeans, modularity, LOF, etc.) And in the book I warn the reader to not use Drive. Slow, clunky, and the Solver implementation is garbage. Plug:

http://www.amazon.com/Data-Smart-Science-Transform-Informati...

Chrome displays a phishing warning for this site.

For me as well

Thanks for reporting it. The site is just a Tumblr blog. The warning might be triggered because of the Mailchimp sign up form.

Ironic. Isn't that the prime directive of Google?

> Speed as a core feature

This is admirable, however slightly myopic. I'd be interested to see how a product can both be fast and have all of the other extremely complicated calculations that Excel has today. People don't use Excel just because it's fast.

You're right, we'll have to make some trade-offs in favour of speed.

You're assuming the product is slow because of the calculations and not because of UI implementation. You're also assuming that the calculations are extremely complicated, by todays standards. People are doing pretty much the same calculations they did 20 years ago, on far faster machines; my laptop is faster than the supercomputer I worked on in the early 90s, my phone is way faster than my desktop was then. It's not unreasonable to ask for the calculations to be done on these machines.

I'd love a alternative to Google Spreadsheets that can handle larger data changes more quickly.

I use collaborative spreadsheets for everything: scheduling, crm, data-loading tasking, actual data entry. If a project is in its infancy and I want to get started quickly for proof of concept, online spreadsheets are absolutely invaluable in all sorts of applications. I piece-meal this stuff out to other services down the line.

I too muck with large tables in a standalone application (libre) and upload my updates. would love to skip that step and stay in a web-browser.

Yes, I do the same and have tried to use Gdocs for this for a long time and I really am sick of it. Please sign up and we will try not to disappoint you.

already done.

If the system can provide linkage to database.quite good.Sometimes customer like to enter data in excel mode aka with dropdown option instead of normal form.

We really want to be able to do that but going cautiously lest we end up like DabbleDB.

What do you mean?

A database with a spreadsheet front-end would be incredibly useful. We would like to try and build something like that. But it's also a hard product to market because for every vertical use case, there are better targetted tools than a generic spreadsheet-db app. See this Quora discussion for some more info on DabbleDB - http://www.quora.com/Why-did-Dabble-DB-never-get-any-tractio...

I think the ideal route might be to grow into that product instead of trying to build it from day one.

Are you considering using the Google Drive SDK and/or their real time collab? Or are you going to avoid any kind of RT in the interests of speed.

I only ask because the app alone sounds like a nice idea, but the key thing is what you integrate it with. Having yet another web app provider that is storing data is a no-no for a lot of people.

No, we won't use any of Google Drive but we will have real-time collaboration. Of course, we don't want to lock-in people's data, so we'll have 2-way data flow between Google docs and other formats.

Why do I get a warning about it being a phishing site when I try to load it?

The fivetran.com has something like that.

I never really stressed it, but looks promissing to big data.

I'm still looking for a really good alternative, exporting my gdocs to excel data processing.

What about Numbers for iCloud? (not a troll question)

The OP's project sounds like it will be web-based. I look forward to the day when I don't need a specific operating system or device to use productive software. Fortunately, I'm 99% there with the things I use.

Numbers for iCloud is web based. https://www.apple.com/iwork-for-icloud/

I see i[Thing] and assume it requires an Apple OS. It looks like this is still the case here: https://www.apple.com/icloud/setup/pc.html

>"To enable iCloud on your Windows PC (Windows 7 or 8 required), first set up iCloud on your other devices, then install the iCloud Control Panel for Windows."

The "first set up iCloud on your other devices" step appears to need an Apple device.

combataircraft would link to http://editgrid.com

Another good example why you never want to use a web service company for important things. Hopefully they will open source their stuff after shutdown.

you all should be checking out ethercalc, i load 30Mb+ csvs into it everyday

Reply on news.ycombinator.com