The world needs a better spreadsheet

Eons ago, I worked for Lotus Software (way before they belonged to IBM) on an exiting new product - their flagship Lotus 1-2-3 spreadsheet product, in those days the preeminent spreadsheet offering in the world, ported to the IBM mainframe.<p>I travelled around Europe as we sold this to large corporations who wanted the power of the spreadsheet, but multipled by a gazillion times and made multi-user.<p>Almost every customer was so receptive and excited, it was a fun and job and easy sell.<p>But later, we learnt that those customers were in fact the early adopters, and Lotus never did manage to cross the chasm with that product. Later Microsoft produced Excel and..well the rest is history.<p>My conclusion: the world doesn&#x27;t need a better spreadsheet. Existing spreadsheet technologies are good enough for the hokey, half-baked things that people like to build with them.

<i>Perhaps, In the end, the best piece of software is the one you never had to build in the first place.</i> -- The Article<p><i>Is it possible that software is not like anything else, that it is meant to be discarded: that the whole point is to always see it as a soap bubble?</i> -- SICP<p>I think the real uncanny valley here lies in maintaining software, not designing it. Business needs change over time and software must adapt.<p>I think the best option is to build your own software. Understand it top to bottom and make it do exactly what you need to do. Sadly this is not possible for most people.<p>The next best option, however, is not to spend $3,000 on something written over a weekend that won&#x27;t be supported down the road. The best option is to go with the Salesforce solution.<p>My first programming gig (in high school) was automating a process that involved manipulating index cards and doing some basic math. I also trained them on how to operate, maintain and extend the system. The end result worked fine for a few years, until it was eventually replaced with some elaborate proprietary system that cost about as much every 3 months as the whole system I delivered. But that system came with ongoing support that obviated the need for any in-house expertise. So they felt the additional expense was worthwhile. And I was happy not to have to take the support calls.<p>Edit: And ultimately both systems were more accurate and saved time over the manual process.

I have been reading HN for more than 3 years. It is like an addiction. I have never felt smart enough to comment something but honestly, this is the worst piece of literature I have ever read.<p>The writer is just rambling on irrelevant stuff while trying to be smart and promote himself and his company.<p>It&#x27;s not even good marketing.

I&#x27;m seeing a lot of hate for spreadsheets here.<p>First, to state the obvious: in the time your least-favourite Excel sheet took to grow from simple time-saver to sprawling Cthuloid monster, it has probably saved <i>multiple person-years of effort</i>. You could spend the next six months rebuilding it all &quot;properly&quot;, and it&#x27;s still comically positive ROI.<p>Spreadsheets are also a unique, top-level category in computing as a whole. By my count, there have only ever been three schemes of interaction with computers (with any significant adoption):<p>1. Shrink-wrap. Use the software you&#x27;ve been given. Read the manual; that&#x27;s what it does, no more.<p>2. Programming language. It&#x27;s text, it&#x27;s got syntax, it&#x27;s got the same basic constructs as any other programming language. You know they&#x27;re all the same, because once you know a few languages, you can start using a new one in an hour or two.<p>3. Spreadsheets. Visual interaction, scaling smoothly from &quot;simple calculator&quot; to &quot;this is the backbone of our whole business&quot;.<p>Spreadsheets really are <i>that</i> fundamental a discovery in the field of computing. Give them some love.

Spreadsheets are the poor man&#x27;s data processing pipeline; they ship with a wealth of numeric and data munging functions and give you an easy-to-understand visual representation of your datastore. You can use them to crunch numbers and graph, but also to implement fairly complex business logic, and they accomplish this within the confines of an application you probably have already bought anyway. They&#x27;re the epitome of the entrepreneurial spirit you need to survive adversity in the SMB space.<p>Spreadsheets don&#x27;t succeed outside of pure numerical calculations because they&#x27;re <i>that good</i>, they succeed because they&#x27;re versatile and ubiquitous, and everyone already has Excel or Google Sheets. If <i>everyone</i> already had, say, Sharepoint, or even some drag-and-drop no-cost rapid application generator that requires zero sysadmin skills to run, that would be the new baseline.

I agree with this in premise, and had a very similar idea with a friend earlier this year - that many of the business apps out there are really just CRUD apps.<p>The main piece of feedback that we got when pitching this idea was that you&#x27;re going to end up as a master-of-none. Yes, we could give you the tools to set up Customer, Ticket, etc. objects. But in order to win over Zendesk, you&#x27;re going to need all of the features they provide. Chat, ticket queues, automation, etc. Business users also don&#x27;t want to be architects - a turnkey solution that solves 80% of their use case is better than a blank slate that they have to start thinking about schemas, relationships, etc. The old saying &quot;nobody ever got fired choosing IBM&quot; can now just as easily be applied to VP of Sales picking Salesforce.<p>That being said, there ARE tools to do this: Quickbase (enterprise) and Airtable come to mind. I&#x27;m curious if the author has looked into using one as their base instead of Google Forms&#x2F;Sheets.

&quot;The world needs a better spreadsheet&quot;<p>I read this when I was a teen (as that&#x27;s long ago). It was at a time Lotus was going to launch Improv in 1993! Improv was the attempt made by a very smart team at Lotus to reconquer the crowd that went to Excel.<p>But as you can guess, the great product never found its market (or never grew beyond a niche) and was killed a few years later.<p>cf. <a href="https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Lotus_Improv" rel="nofollow">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Lotus_Improv</a>

While I&#x27;m not entirely sure if &quot;the world needs a better spreadsheet&quot; is a stand in for &quot;accessible prototyping tools&quot; it&#x27;s worth noting that things like Quantrix Modeler and DADiSP are the better spreadsheets it&#x27;s just that no one uses them because Excel (and by extension Excel work-alikes) have path dependency on their side.

Fundamentally software development is about writing and managing complex systems. Excel and the like have limited capability to manage complexity and any simple system built with excel eventually turns to bad smelling unmaintainable chaos. So then people build tools that manage complexity better but are still &quot;easy to use&quot;. And then those tools become so complex that software developers are the only people who understand how to use them. And then what&#x27;s the difference between said tool and writing real code? After 10-20 years we are back to where we started. I wonder exactly how many times I will see this in my career?

Not really. Most spreadsheets(~88%) have errors.<p>&quot;Sorry, Your Spreadsheet Has Errors (Almost 90% Do)&quot;<p><a href="https:&#x2F;&#x2F;www.salesforce.com&#x2F;blog&#x2F;2014&#x2F;09&#x2F;how-to-reduce-spreadsheet-errors.html" rel="nofollow">https:&#x2F;&#x2F;www.salesforce.com&#x2F;blog&#x2F;2014&#x2F;09&#x2F;how-to-reduce-spread...</a><p>The 2008 source article is<p>&quot;What We Know About Spreadsheet Errors&quot;<p><a href="http:&#x2F;&#x2F;panko.shidler.hawaii.edu&#x2F;SSR&#x2F;Mypapers&#x2F;whatknow.htm" rel="nofollow">http:&#x2F;&#x2F;panko.shidler.hawaii.edu&#x2F;SSR&#x2F;Mypapers&#x2F;whatknow.htm</a>

It exists already. It&#x27;s called MS Access.

I agree, I automate generation of hundreds of spreadsheets a day.<p>I would like to see more scripting&#x2F;automation features on email.

Quip (a better spreadsheet) or Django web framework (its ORM and built-in admin can do wonders) are worth mentioning. Another example that comes to mind is how Vox uses Google Sheets for content &amp; config: <a href="https:&#x2F;&#x2F;product.voxmedia.com&#x2F;2014&#x2F;7&#x2F;29&#x2F;5863004&#x2F;take-a-peek-at-the-code-that-powered-the-verge-50" rel="nofollow">https:&#x2F;&#x2F;product.voxmedia.com&#x2F;2014&#x2F;7&#x2F;29&#x2F;5863004&#x2F;take-a-peek-a...</a>

I built an app in Excel 5 years ago. It was a heat map for events, for stadiums, theaters, etc. I would build the static CSS&#x2F;HTML and write out the files from Excel. It is still in use today. After having stopped working on it for a couple of years a minor bug was brought up to me recently. Luckily I had a windows computer again with Excel installed. Excel is a great platform for building prototype apps, or even long lived apps - especially 10 years ago. It is amazing the power that Excel has. But, a proper app should be built in JavaScript with a solid back end (F# being my fave). Although developing in Excel was a bit clunky it was amazingly easy to get beautiful customer interfaces and beautiful forms.

Being part of design meetings, it&#x27;s amazing how many endless customizations and workflow adjustments people like project managers and consultants like to request for a new piece of software. But when it comes down to it, most users don&#x27;t want a complex, hard to understand interface that has so many features it&#x27;s supposed to improve everyone&#x27;s productivity. Most people want to use their spreadsheet, their Outlook calendar and their IE browser. Not because it&#x27;s best at the task, but because that&#x27;s what they know and are used to.

Once you get all the security, backups, data validation and whatever exporting&#x2F;search capability the people need a system like this is never simple. I wonder how happy they really were with his software vs relieved to have a semi-functioning system.<p>It still seems like something like this is in the realm of possible. Salesforce mostly does this, but its UI is garbage and most people end up hiring implementors for it.

Quickbase does an amazing job with allowing non-tech users to create their own crud apps, and the model is very spreadsheet like, including grid edit if you want.<p>The trouble is the pricing model. $25&#x2F;user per month for the edition that&#x27;s not hobbled. For some use cases, that makes sense. For others, it&#x27;s just way too high.

Airtable is a better spreadsheet imho.

What&#x27;s with that garbage at the bottom right? I&#x27;m trying to read your article. Why would I want to talk to you before reading your article? Made me stop reading.

Charged a close friend $3,000 for a few days worth of work?

how do people get a &quot;hey HN&quot; welcome line up on their site so quickly? This was posted here less than 2 hours ago. Do people really monitor incoming referrals hourly?

I swear to god, anytime anything pops up on a website when I click on it, it seriously pisses me off.<p>A fucking chatbox? I already hate your company.<p>A fucking chatbox from the author of a blog? I hate you and everything you stand for. You are what&#x27;s wrong with the internet. You should know better, and you do. But you do this anyway. Just stop.