How do you build web applications?
My partner and I are trying to understand how developers/makers build web applications.
Do you start development from scratch? Or would you use a Low-code development platform?
When previously building applications, we’ve experienced a lot of repetition which annoyed the crap out of us.
So we have spent the last year building a SaaS builder - Budibase - https://www.budibase.com/
Budibase is a platform which eliminates repetition, speeds up development, and allows you the freedom to use your own UI. We’re also going to open source it.
The product is not launched yet but we thought we’d seek validation from the Indie Hacker community if they feel this is a good idea or not. We aim to go to market in the summer.
Is this something you would consider when building your next application?
Your answers are appreciated and will help us decide the future of Budibase.
Cheers and happy new year!
How does Budibase compare to Bubble.is? I use Bubble for everything, partially because I don't code well, but mainly because I can get to the valuable part of whatever I'm building super fast.
Thanks for your response Andrew.
We are targeting the same market as Bubble.is - so similar. We feel that we will be more customisable than bubble.is - able to get deeper into the code, plugins and customisations when necesssary. We will be open source from day 1.
Sign up on our website and we'll let you know when we launch. It would be great to catch up and ask your opinion on the product.
I'll definitely sign up. I teach courses and workshops on building apps on Bubble, and am one of the moderators on their forum. I'm heavily involved in the low-code/no-code movement, and eager to help however I can.
Andrew, that's amazing. We would really appreciate your knowledge going forward. We'll be in touch. Happy new year!
What is the technology stack?
All javascript. Nodejs web server. For front end, we will be framework agnostic, as we want anyone to use whatever framework they like to write custom ui. Currently, the only library dependency of the core framework is lodash.
It will also be mostly Database agnostic... any key value or document store can be used (including file storage). Our hosted version will use Azure Blob Storage
Also, thanks for the question Gates, appreciate it.
In general I use Rails. I'm a full stack developer and can get some pretty complex sites up and running quickly. I HAVE been recently looking around at using a service for generating more simple sites. Things like marketing landing pages, email catchers, etc. One that I have been looking at recently is https://www.launchaco.com/. It seems budibase is targeted at the SaaS market, not sure if I'm the target.
Thanks Patrick. I do not have any experience with launchaco, but we would not describe ourselves as a website builder, like they do. The output will be a single page application (and/or mobile app), with structured data storage, and some common functions on top (such as user management, user roles, search etc..)
I usually use heavy lifting web frameworks to build a projects.
Since I'm comfortable with my tools, I can go from an idea to MVP in matter of days.
Tools I use: Python/Django, TypeScript/Angular, Ionic (mobile)
Great work Alireza,
We use javascript. Nodejs web server. For front end, we will be framework agnostic, as we want anyone to use whatever framework they like to write custom ui. Currently, the only library dependency of the core framework is lodash.
It will also be mostly Database agnostic... any key value or document store can be used (including file storage). Our hosted version will use Azure Blob Storage.
We want to speed up the development process, hopefully by 30-70% (estimation at this point). We also want to reduce the pains within the development process.
Visit our website, and sign up to Budibase and we'll send you a quick message when we launch.
Have a great day.
Regardless of the language, most devs probably use a framework in addition to a template with basics such as authentication. Then they pile on language or framework plugins.
What makes Budibase different?
imo Something that further automates the status quo would be a very specific base sample app i.e store, social network. Is that what budibase aims to be?
Your landing page is so vague that is seems that you're just garnering interest at this point, instead of having anything working. Nothing wrong with that, but that's just my impression.
I agree with your last point. Too vague landing page.
The landing page should describe how the apps are built, is it similar to SquareSpace ?
The landing page is vague. We understand this and we plan to develop it with more content in the next few weeks. We first wanted to gather feedback on whether or not visitors could quickly grasp what we are trying to do.
Out of curiosity, do you prefer technical content in writing or video format, or both?
I haven't used Squarespace - but i think the sentiment will be the same (except applied to a data-storing web application) - using a UI to build a single page web app. You will be able to write code (in javascript) to pull data in and out of your application, as well as your own custom UI
Also, I really appreciate your comments.
With budibase, we aim to cover... Data Model design, Data Validation Rules, Search, Data Object Relationships, User Management include Configurable user levels/roles, Data aggregation (stats like sums, counts etc)... all this configured using a user interface.
Then, we want to be able to generate no-code UI for all this - but allowing you to add your own UI where you need it, using the framework of your choice.
In terms of progress... we have about 90% of the core framework completed ... but we need to wrap this in a web server, and do all the UI... so we are close to 50% of MVP.
When I saw this SAAS, I remembered your service.
https://bullettrain.co/
Big commodity features already done out of the box
Thanks for remembering Chaos. Bullettrain is an interesting product. It's pricing is also interesting.
We also aim to have the same features and more :-)
If we had to position ourselves it would be somewhere between Bubble and Firebase.
Great spot and thanks again.
This comment was deleted 8 years ago
Personally, I start from scratch but would appreciate less repetition and a slight head start. I've often thought of building something like this but never had the time. Good luck!
Cheers!
Does anyone have any experience with headless or API-first content management systems (CMS) to power the backend? Including databases, user authentication etc.
I've dabbled with Drupal and Wordpress but have found them limiting when it came to data structures/relationships and frontend UI design.
I've come across offerings like Contentful and ButterCMS but don't know much about them yet. I'd love to hear some experiences or tips, or recommendations especially free or open-source solutions.
In my case, when I conceptualize the bare bones of my product idea, is that it ultimately can run with a CMS, hence my initial interest as the preferred development direction. I'm not a coder, so the CMS could do the heavy lifting, only remaining the UI design to sort out (ideally a web app, or PWA).
Thanks everyone
Hi guys,
I'm technical guy and from my perspective all web builders are not time savers.
Let me explain, when you build your first product by coding you spend a lot of time for sure, but you should consider it as in investment for future projects. If code is well organised and decomposed it has reusable parts such as authentication, data storage, configured infrastructure (docker, etc).
So when it's time to build a new project you just combining these parts together quickly, changing CSS/fonts a little bit, writing basic back-end logic and that's it.
Web-builders are trying to reach this speed of development but they are always behind of building product by writing code.
Another problem is that web builders have an illusion of flexibility. If you need a really sophisticated piece of logic the only way is to code it yourself.
What I really need from such apps it's having possibility to change code generated by web builder to customize it as I want.
Hope it was useful!
Thanks Eugeny.
Your points are well taken about web builders. I have often found that I end up getting my hands dirty with code, despite starting with a web builder. Even recently, when using Create React App, the reactjs boiler plate, I have never actually been able to complete a project without ejecting from Create React App ("ejecting" = forking from the boilerplate path).
However.. I see the front end UI generator as a relatively minor and less important aspect for BudiBase. Anyone creating their own SaaS app MUST be able to customise their front end completely.
The BudiBase backend are where I see the big savings.
... all just off the top of my head
Many of these things are repeating across projects
Of course, a big challenge for us will be to make it pluggable / customisable enough to allow unique applicatione to be built. But, I believe we can do it
Thanks for your comment, much appreciated.
I'm a from-scratch Javascript and node.js person. Client side I don't use frameworks unless you count my own libraries. And I only use 3rd party libraries for very site-specific things. Server side I use Express and Postgres, and typically host stuff on Heroku.
Budibase is built with javascript. Nodejs web server. For front end, we will be framework agnostic, as we want anyone to use whatever framework they like to write custom ui. Currently, the only library dependency of the core framework is lodash.
It will also be mostly Database agnostic... any key value or document store can be used (including file storage). Our hosted version will use Azure Blob Storage.
You should sign up to get notified when Budibase is launched. You can do it on the website - https://wwwbudibase.com . We would appreciate your feedback.
Have a great day.
A great book written by Brian Kernighan and Rob Pike (both computing luminaries), The Practice of Programming, suggested that one of the three critical pieces of programming effectively is automation. I tend to agree. If you have a lot of repetition, then automate it, script it or create a macro to do it and similar tasks.
Using a rapid application development framework (like Rails/Laravel/Phoenix, etc) is a good idea, as is using the highest level language you can. But areas a lot of people ignore are automation and truly learning / optimizing their editors. If there's any kind of boiler plate you're doing again and again, you should set up a way to generate it in in a few keys.
Thank you Alchemist for the book recommendation and your comments. I feel we are definitely on the same page.
Sign up to Budibase and we'll let you know when it is launched. We would really appreciate your feedback.
Have a great day.
I wouldn't ever choose JS on the back-end again, especially for an indie project. I worked with it professionally for years but later discovered more productive back-ends.
You can't do much about the language itself, its weak typing or its poor stack traces, but you can copy some of what works from widely used MVC frameworks. Make generators for controllers and templates, handle database migrations and make clear default options. That would be my feedback.
Hi, thanks for your advice, previous experience always welcome.
I sort of agree with your view point on Nodejs as a backend. I am a .NET dev by trade and would almost always choose C# or F# for a backend. I enjoy strongly typed, and the tooling is second to none.
However, I have my reasons! ....
This is a very dynamic project - a platform for building apps. The dynamic typing really really helps here in many areas. I actually started with Typescript (library needs to be used client side) and found that types were massively getting in the way.
It is saving me alot of time being able to use the core library on front end and back end. For example, executing custom data validation rules - I only have to parse these in one code base, and can use them to validate data on forms in the UI, and as they arrive into the HTTP endpoints. This is one example of many point of reuse.
Thanks so much for taking the time to share your experience, but I'll struggle on with good ol' Javascript!
In regards to dynamic vs static typing, I'm somewhat ambivalent. Many languages I'd choose over JS are dynamically, but strongly typed (e.g. Ruby, Python, Erlang, etc...)
Javascript loses on a number of fronts. From the indiehacker's perspective, productivity is the largest issue, but ease of debugging, concurrency, scalability and fault tolerance are also major issues as you grow.
Validations need to be done on the back-end. Front-end validations are unnecessary and mostly serve as a mechenism of reducing requests to the server (an optimization far less important than speed of development for indies!)
Setting all of this aside, if you want to win, I'd suggest putting every possible effort into being more productive than Sails JS (and maybe even Rails, eventually).
Static typing is what i meant, not Strong typing! Thank you, I always mix the two up.
Your point on Rails/Sails - I haven't really considered us to be in competition with these frameworks. However, now that you mention it, that seems obvious. So, thank you for the insight!
I am completely expecting BudiBase to be much more productive than both of these, but this will be a great yardstick for us (although will be hard to measure).
Thanks!
For my current product https://www.indiehackers.com/forum/private-contact-and-relationship-manager-with-client-side-encryption-8d26d1d142 I setup everything from scratch. I am not sure if a SaaS builder would help me with this type of app.
For other project, I'd like to get rid of server setup & repetition, but I am worried about
Hi Petr, thanks for the feedback.
To your points:
The goals for pricing are
Basically, it has to make sense for all parties.
Of course, if a customer were to out-grow us, it might make sense for them to run and host the platform themselves - which we wholeheartedly support - its good for the brand.
We will be trying to make it as plugin-able as possible - meaning that it should be easy to write other services in technologies of your choice, and integrate using the Budibase API.
It's also worth noting that the database used by Budibase will also be pluggable... any key-value or document store is perfect... and can also work with file storage. Our hosted version will (probably) use Azure Blob Storage (similar to Amazon Dynamo DB, i believe)
Our hosted version will of course store your data. The Budibase APIs should be sufficient to extract that data at any time, should you want to self host.
Thanks for your feedback, its been helpful for me just to type these ideas out!
As a Designer and a semi-amateur coder, I kinda want the best of both worlds. Common drawbacks I see with sass web builders is the lack of customization & support. For me, the sweet spot would be a Sass that gets you up and running quickly but also offers customizations later down the line when your app grows.
Currently, I use React (yes with all the bundles) and Firebase for Single Page Applications and WordPress for portfolio sites & forums.
Thanks for your comment. That "Sweet spot" you talk about is exactly what we are aiming for.
Most web apps have 80% of functionality in common. It's that last 20% that make's your application unique. We aim to cover the 80% and allow for the 20% customisation.
Completely fabricated numbers there of course.
(BudiBase partner here)
Hello from a fellow startup founder in Belfast! Happy to follow and support another local. I'll signup to keep track, and because we know it's an important metric when in conversation with anyone for the social proof. :) Good luck folks!
Amazin Lee! Great to see some local eyes.
We would love to connect/meetup and get some of your experience, we'll take all the help & advice we can get.
So my concerns with any PaaS like this is lock-in and dependency and can it scale with me. Many times my MVPs are something I plan to throw away (the code) if they are successful. They aren't meant for scale.
Thanks for your comment Brian.
I fully agree with your point on lock-in. We aim to be open source from launch day. So, you will be able to download, self host and modify the source code as you need.
We aim to license the core framework as MPL - meaning that you must open source any changes to the core library. However, your own apps and customisations are your own.
In terms of scale - we aim to run our own hosted version almost entirely serverless - so scale is of primary concern from day one. In terms of scaling the data, we are looking at databases records into the millions.... so not big data, but enough for most applications.
I think the challenge is also how do you build a truly flexible and easy to use platform. Also take inspiration from airtable and explore their universe section to get ideas on how to inspire others on what they can do with your platform. Sell this on solutions not features. TONS of people don't care about the stack, just want something that gets the job done (jobs to be done framework).
Airtable is great, we can pull a lot of inspiration from them. Budibase, at the beginning will require a small amount of coding knowledge, but will offer a lot more flexibility and potential.
I complete agree we need to sell solutions. What's that old saying, 'features tell, benefits sell!'
Before we go to market, we are going to build a few templates/products so people can see its potential.
Documentation is also crucial. We hope to write documentation and provide video tutorials.
Cheers for your feedback Brian. We'd love if you could signup to Budibase to receive a notification when launched. We'd also love the opportunity to chat further when the product is more developed.
Have a great day!
Interesting. Currently I am building up with Django. If for whatever reason this project doesn't work out I will reuse the code and just pull out my app. I wish I knew about Firebase and all this other stuff you peeps are talking about. Just found out about this stuff a few days ago, but looks awesome for rapid validation.
Django is great Ryan. Budibase focuses on Javascript. I hope your app is coming along well.
Building apps can be tricky, hence why we created Budibase. We're trying to take the pain away.
Signup to Budibase using the contact form and I'll send you an email when it's launched.
Appreciate the feedback!
Also a more technical answer, if required:
We will be going an extra few steps than DJango. The datamodel, data validation rules, user management (including user roles/levels), data indexing will all be manageable via a UI. You will be able to create a working app without writing any code - but thats not really the main point... the point is that you can quickly produce the basic features that all apps need then code the really custom bits that apply only to your idea. Django is giving you a framework to build your app upon. Budibase is giving you your basically functioning app to build upon.
When I need to build a quick sprint demo or MVP, I'll use tech tools that allow me to build quickly.
Tools such as front-end and back-end framworks. Even use database such as Firebase to handle data & user authentication.
in just a few weeks, a MVP can be build and tested to see if it has value to customers or whoever.
Thank you yNinja. We really appreciated your feedback. We were the same, always developing from scratch,but we found there to be a lot of repetition. This is why we created Budibase. Would you ever consider using a a tool like Budibase?
yea for sure. Just signed up.
Thank you. Cheers.
I use a similar boilerplate I created, I never build from scratch.
Budibase looks cool, although you're website doesn't explain how it works from a technical point of view. Is a framework that can be downloaded? How does that work with your built-in hosting?
Thank you Kyle. We, agree more technical content is required for the website. We purposely provided as little info as possible so we could receive feedback like yours. We really appreciate it.
Building from scratch takes a lot of time and involves repetition which was the encouraged us to build Budibase.
Budibase is similar to WordPress. Self-hosted, or hosted by us.
But it also handles your data-model and storage and all validation and rules that go with it.
At its core, it's a framework, so can also be used in a client only fashion... e.g. desktop or mobile app storing to local file store or G Drive.
The plan is to develop the code in private until we launch V1. We want to completely control the direction of the product to V1, before having to manage a community.
However, I am not totally against opening it up for people to have a poke around. Do you think there is benefit to that at this stage ?
Appreciate the feedback.
We don't think 'hordes' but with the team we have at hand and resource available, we hope to launch a marketing push prior to product launch.
This comment was deleted 8 years ago
This comment was deleted 8 years ago
Thanks for the feedback. I totally agree - docs is the first place I would look, we just haven't got that far yet.
The basic idea is that you would use a builder UI to create a functioning app, that has all the standard, repetitive functionality that all apps have. Then, you add customisations (UI and backend) on top - which really are the things that make your application special.
(BudiBase partner here)
This comment was deleted 8 years ago