I am so excited for this software. I have used Airtable extensively, and it's a pile of garbage. Here are just a FEW of the "features" Airtable charges $24/user/mo for:
- Reliable 1-5 hours a month of downtime. There are months where it's literally 1 nine of uptime. Their status page was actually down for a while (I think they forgot to renew it?)
- Simple and obvious bugs in the official API clients have not been fixed for years (for example, Promises in the official JS library throw the string `[object Object]` for all exceptions.)
- The API has no way to query the structure of a table, you have to infer it from the first few records. (The auto-generated API documentation is actually wrong if you don't have a fully-filled-in row somewhere in the first few rows.)
- You can't modify a table configuration by API... except you can _sometimes_ by adding a parameter which is documented to try to convert dates, but actually modifies field definitions for a number of fields. Documentation for which fields is entirely in scattered forum posts from users.
- Booleans in the API are `true` or `undefined`. Seriously.
This is all /barely/ worth it because Google Docs doesn't work when you need a large structured database with many collaborators. But man is it bad software.
When you’re doing dynamic structure in relational dbs (traditionally) options include storing a wide row with all the data types and you choose which column to use, or storing as lowest common denominator (string) and casting in and out at runtime. These days you can work around this stuff with json / schemaless models.
Airtable allow switching data types on a column (from memory). Maybe they store everything as string and then the column definition is for display and allowed operations.
Hubspot do this (poorly) so they’re forever giving you back things like dates as string of a number representing a date.
Try force.com (fully featured developer account is free with limits), very stable and fully featured PaaS with long tail of features. On a quick review Airtable as a slick/modern UX on less sound architecture than force.com, probably shipping features in a breakneck pace.
Not a Joke. I am recommending Salesforce.com's PaaS offering (used to licensed as force.com, now called "platform" [0]). Expensive, but it is IMHO the best low code cloud platform there is that has got every aspect (Apps, Identity, Security, ACLs, Workflows and Triggers, APIs) right.
> sanity as a developer
Low code is targeted at app building experience for non-developers/citizen coders. Curious if you have personal experience with the Salesforce platform and can provide details on some of your biggest gripes with it.
I have used it with great success in building simple “textbox on database” applications that are easy to maintain and change.
> hearing Salesforce alone
What do you think about heroku?
I am genuinely interested in alternatives that is cheaper and don’t charge per user.
I have never used a Salesforce app that was not terrible. I'm sure it provides lots of value to the business who use it, but it is awful awful for end users it is thrust upon.
My experience with Salesforce is exactly the promise of low code. Because you can get a lot more done without needing to be a software engineer, it empowers a whole extra set of people to build software without having to think in terms of reusability, performance, scalability, or a coherent data model. And you see the results of that.
I think as a platform it’s really powerful. (I’ve arrived at that after years of looking at it with disdain). But it’s problematic from our perspective exactly because it throws out anything we consider good habits of software development.
Force.com used to be a dedicated domain for salesforce's easy to use enterprise development platform, which is quite popular. Seems like they've rolled it back under salesforce.com as a page.
Just curious - what do you use airtable for exactly?
When do you need a structured database with many collaborators?
Airtable and the like strike me as a database tool for dummies who want 'anyone change anything anytime', which is just a recipe for no one knowing who the hell is changing what and when anymore.
Microsoft Access and Lotus Notes (and to a lesser extent SharePoint) both had these characteristics and are quite popular for the long tail of business processes that aren't covered by dedicated vendors. The only problem is the tech tends to age and it becomes hard to migrate the 1000+ bespoke "apps" that live inside the enterprise.
"anyone change anything anytime" is much better than "hey can you email me that excel file? I'll make a few changes and then email it back to you" "actually I don't have the latest version, Person X has it and they're currently merging someone else's changes, so be patient"
It’s my understanding that the main selling feature of AirTable is the hard coded views that they offer; calendars, kanban style planners, things like that. The slightly more structured nature of relational tables over spreadsheets lets you implement those fairly reliably.
I use the free tier as a collaborative Kanban board for a tiny 5 person team.
Airtable allows me to define arbitrary fields for the Kanban cards, which happens to be a premium/paid feature in all the actual Kanban saas tools (Trello, etc).
JIRA would not just overkill but a bad fit for us, potentially requiring extensive configuration.
We're an academic research lab that also happens to do contractual work for other labs (mostly internal to the university), mostly designing and building mechanical prototypes and testbeds for other researchers.
It took like 5 minutes to set up an airtable base to track our WIP and all information we use for billing our "customers": hours, material purchased, internal budget reference, etc.
Yeah, most SaaS solutions have this problem. GitLab wants you to pay $80/user/mo to be able to get dependency scanning or vulnerability scanning. While Github gives you that for free. They could at least include it in their $20/user/month tier.
At some point you have to pay a price to maintain a competition in the ecosystem. We all know what happens when there is only a player left. Microsoft can just sink billions in side games like Teams and GitHub at the moment.
Personally, I think these two features should be part of the $20 tier and not the most expensive one. I don't understand why I have to pay $60/user/month more to get some UI integration which I can also get when paying for GitHub Enterprise for similar price as GitLab's $20 tier
- Reliable 1-5 hours a month of downtime. There are months where it's literally 1 nine of uptime. Their status page was actually down for a while (I think they forgot to renew it?)
- Simple and obvious bugs in the official API clients have not been fixed for years (for example, Promises in the official JS library throw the string `[object Object]` for all exceptions.)
- The API has no way to query the structure of a table, you have to infer it from the first few records. (The auto-generated API documentation is actually wrong if you don't have a fully-filled-in row somewhere in the first few rows.)
- You can't modify a table configuration by API... except you can _sometimes_ by adding a parameter which is documented to try to convert dates, but actually modifies field definitions for a number of fields. Documentation for which fields is entirely in scattered forum posts from users.
- Booleans in the API are `true` or `undefined`. Seriously.
This is all /barely/ worth it because Google Docs doesn't work when you need a large structured database with many collaborators. But man is it bad software.