Ask HN: Is it a waste of time to code internal tools?
I think I just developed a prototype that can build internal tools automatically (to a great extent), for any company, and has inherent support for dealing with certain entities such as users, cards, payments etc., As easy to set up as Stripe, and as easy to use as Zapier. All of this instantly, with minimal coding (only if you want to change the visual appearance of the tools generated & add custom logic using JS & Python). Also, there are efforts to integrate this product with other tools like Salesforce, ZenDesk, Kissmetrics etc.,
This product will reduce the time spent in developing internal tools, and the engineers in these teams could be used to build something even better.
So the question here is: Is it a waste of time to code internal tools manually? What are the issues you frequently encounter in your companies w.r.t internal tools (development & usage).
Also, is anyone interested to be an early adapter and use our product to generate internal tools for your companies/products (for free)?
I think it's highly unlikely that we're talking about the same thing here.
I didn't mean the in-house tools we build to make our dev's work easier, but to make it easier for the non-dev members of the company.
I was referring to all those web UIs we build in companies for our internal tools to use to service our customers, help our sales guys achieve their targets etc., In other words, we could just think of them as an UI with the production DB's slave in the backend which allows the internal teams to consume/edit data. Think of user management tools, log analysis tools, sales trackers, fraud trackers etc.,
For small businesses there's definitely a market for automation platforms but there's a lot of problems in actually selling these products. You either focus on a key area where they already have a solution they don't question or you go too broad and confuse the people you're selling to. Like let's be honest, your tool can't build everything.
I think the question you gotta ask is why is your tool builder better than the frameworks and resources that devs already have available to them.
Internal tools can be just another tool that makes people's lives easier. They can be an albatross when they become someone's little fiefdom, or they can become Slack. You never know where they can go, but the bigger questions before I go building my own:
Is there anything that already gets me close? Usually
Can we change how we're doing things to better fit something that exists?
e.g. Slack started as an internal messaging tool at a game development company
- "The Mythical Man Month" p 128, "Sharp Tools".
To answer your question, it depends. Anecdotally, we created custom solution for analytics for our Robot, and then moved it in as a solution. You can think of it like 'internal-pivot'.
Additionally, in-house tools, if designed correctly, are usually small pieces of software that do just what a company needs. Tailored solutions, if effectively created, are sometimes better than something out of the box. It also lets the developers stay stimulated when there is not 'actual work' to be done and adds potential for features to be implemented for their product; if not providing with something they can pivot and sell separately.
But briefly, here is how it works. 1. You give the schema of your DB to the tool. 2. The tool analyzes and build backend flows. 3. The user can now select which fields indicate the username, card or payment amount etc., and the flows will be adjusted for the same. 4. User can add access control to tables and fields. 5. Preview and publish, and the tool is ready to be used.
Since a single tool like this cannot build tools for specific usecase, there is an option to define own actions and layouts.
The best advice I could give is ensure you're doing it for the right reasons, that everyone around you will truly profit and make sure you've honestly evaluated existing tools.
I often see people failing to correctly evaluate an existing technology/tool and then rush off to build their own inferior product. Yes it can be fun to just code, and there are cases where it's truly a good / beneficial thing to do, but they're not common from my experience.
Good luck!
There's always going to be demand for tools of many kinds, so whatever you're offering has to differentiate better than "internal tool". The abstract you give is kind of like a software framework, possibly with some form of high level scripting, configurator, or wizard to guide the user.
At the end of the day always evaluate your solution against a short bash or Python script. Did your solution gain anything? If not, why?
That's not to say don't try, but be aware that this is harder than it looks!
The problem is, they're still programming languages, but very opinionated They look like magic if you go the way the tool allows you to go, but when you go off the rails you end up wasting more time than if you had coded it in a standard language.
But yes, its definitely harder. Perhaps it seems like I am looking at it vaguely and broadly, but I will narrow it down to a specific type of tool or to a specific type of business.
And excuse me, I dont know what your tool is or does, but there is no way that it could replace even a fraction of this development (though that doesnt mean it wouldnt be useful or profitable). Otherwise it would be intelligent enough to replace developers entirely.
Sadly, companies often end up going with something like salesforce, which I hear also has the ability to be extended like these more general tools (though they try to sell consulting services to do so). The reason companies go with something like salesforce is that 80% of what they need is already built and done in a pretty targeted way. For the rest they can have an in-house dev or a consultant extend it and integrate it with whatever other tools the company is already using.
Irony, I wrote it for AppEngine, didn't host it though.
I would also like to point out that you're asking people whether they want to be an early "adapter" of a product you haven't described or referenced. "I think I just developed" doesn't install any confidence into startup wanting to adopt your mystery software, either you have or you haven't.
Opinion: What you're describing sounds like a large API. I think you may struggle to get adoption if you're looking for money from this software.
I intentionally avoided describing the product as I wanted to know more about the problem space, so I can come up with the right solution using this prototype.
But I just realised that Google is working on something similar: https://developers.google.com/appmaker/
https://developers.google.com/appmaker/
Should I be giving up?
If you think they won't ever fix that, there's an opening there. It's an odd limitation to me. I suppose they are trying to sell GSuite licenses.
Or maybe tie it to AWS with a pre built AMI? There are at least some companies where a point-and-click AWS instance would have value.
Do you think pre-built AMIs might be a good idea? I am not sure about it though.
There are plenty of cases where it doesn't make sense. I am not surprised companies build tools even when it isn't logical.
I would like to try your product, if I find it useful I might be willing to pay.
You are paid to do that, at least it means that worth your compensation, right?
So, a new programming language?
For example, the tool processes the DB schema, builds a UI. Learns which field is username, and then shows in the front end the way user management softwares work. For the sales guy, its just all the information from Salesforce, different forms, the value of their converted leads. For the finance team, its the payments, fraud disputes etc.,
All this data is in the DB. And engineering efforts go in to build these tools for the non-engineers and keep re-iterating. I am trying to solve that problem of re-inventing the wheel. Any functions specific to the business can always be added using JS or Python (just like Zapier).
There's a continuum of automation software, non-programmers often find themselves comfortable writing stuff in Excel, and some become kind-of self-taught programmers using something like VBA, Access, or whatever is used today, but they are still not programmers and don't understand how your "magic" tool works.
If your tool has the ease of use for non-programmers as Excel you might have something big.
If not, you still might have something useful :) but understand your audience :) .
I am guessing it will be as easy as Excel for the non-programmers to use the tools generated using my tool. But, only the users can say.
In fact, we were debating on providing an Excel like interface (which will be auto-created if the table has too many columns, and can be manually created if required) that just boosts productivity when performing data manipulation jobs.
We saw that most non-programmers, when confused, just look for an export option first, so that the data can be analyzed in Excel. Then they modify it in Excel, upload it on to the systems again - so yes, have a tough competition here with Excel, but we are trying to get that easy to use factor into the online interface to make sure the non-programmers won't have to use multiple tools to get a single job done.
But I really hope we come with a product that generates tools that are super-easy to use. Might take time, but that's where we want to go.
I've used most of them, and they end up being frustrating. Quickbase was the most flexible of the lot, but it's expensive. I haven't tried Google's. It sounds like it has a problem many of these tools share though. Per user pricing, with no notion of "external users". Think something like job applicant tracking. You couldn't offer an external user portal to upload a resume without paying a license for every applicant.
We were thinking of per-user pricing only for the employees. The tools generated will have an "Allow external user signups" option that would allow external people to signup, and then the external users' traffic would be charged by the hosting company (or by us if the user chooses us to host their tools for them). I think this is a normal way of doing it, and I believe these tools will soon switch into this mode, if not, I am glad :D