What’s the Difference between a bug and an edge case feature? And why is that important to know when using browser automation?

In the fast-paced world of automation and technology, distinguishing between a bug and an edge case feature is crucial. Join me as we delve into a real-life scenario shared by Jeremy Redman, the founder of Task Magic, shedding light on how a seemingly minor issue can make a significant impact.

Jeremy's narrative begins with a customer's report of a 'cloud issue,' highlighting the complexities of automating web processes. By sharing the experience of a dedicated customer, Hans, who encountered a unique roadblock, Jeremy emphasizes the importance of community-driven solutions.

Through Hans' experience with a scroll down step that overwhelmed the system, Jeremy elucidates the fine line between a bug and an edge case. The intricate details of how a specific action led to a storage limitation offer valuable insights into the technical challenges faced in automation.

The heart of Task Magic lies in its commitment to rapid problem-solving and community engagement. Jeremy's transparent approach to addressing edge cases within a day exemplifies the power of collective input in refining automation tools.

As Task Magic evolves, Jeremy unveils plans for an educational course to equip users with diverse strategies for overcoming site-specific challenges. By encouraging patience and active communication, Task Magic aims to empower users to navigate complexities with confidence.

In closing, Jeremy's recount of the bug-edge case dilemma serves as a testament to Task Magic's dedication to innovation and user-centric solutions. Embracing the unpredictability of edge cases while resolving them swiftly underscores the brand's resilience in pushing boundaries.

Hey everybody, Jeremy Redman here, founder of Task Magic. Give this channel, give this video a like, give this channel a subscribe. If you like automating work life balance, we're going to put, we're going to see the things that you want. To do Kyle just said one second. I was going to do this without him.

He'll he'll call me. All right Okay, call me now that kyle's late So I wanted to go over this one topic and use this one example this one person lovely customer Love this person they've helped and I don't know if they've thought about how much they've helped. Um, okay. So here we go So hey, you're on the pod right now.

Okay Okay, so I and I just started this it's 47 seconds in. Okay. Now I want to, I want to talk about, so, so Kyle knows as well. I want to make this video about the difference between someone interpreting something as a bug and it just being an edge case feature build. Okay. So, and I want to describe that the cloud issue.

Okay. So the thing that was reported to us as a cloud issue. All right. So by the nature of what we do, we are trying to automate the web and see, see a couple of videos ago where we're trying to boil the ocean. All right. So in doing that, we are having to account for everything on every website. And doing every action on every website because we give you the freedom to do that.

We want you to do that. The only way we accomplish this is if you let us know the esoteric sites or, or esoteric or rare edge case things. That isn't happening. So, for instance, we had one of our lovely, lovely customers. I'm going to put this up here. So I'm going to share a screen. So, Hans, one of our customers reported cloud not running again.

OK, so and then he brilliantly put a video. We love you, Hans. Love you, by the way, because this is what we need. So, oh, oh, you just wrote back. Oh, I think cloud is running now. Okay. Yes. Oh, oh, he just put that 17 hours ago. Okay. Yeah. So we fix that. Um, so here's the funny thing that we found out Kyle off memory.

Do you remember what that issue was? Um, it was so we try to generate a description of what every step is for people because like, if you record a click step, we want to be able to describe what that step was back to you. Um for his automation, he recorded a scroll down step on a site that returned so much content Yeah that we didn't have a limit on what we were Interpreting there and when we saved that it was exceeding the storage for the tap Which made no task history show for him.

It was a very weird, it was kind of a, we haven't recorded a scroll down on a site with that type of a situation before. Yeah. So it introduced a new problem, right? So it was a payload issue. Okay. So it was one, a, it wasn't like a commonly used site. Right? Yeah,

I don't think so. Yeah. So, I mean, I think it's like a, it's a more niche site. I don't know. Yeah, yeah, yeah. Right. So, Again, So, that's fine. And our thing will work at least 80 to 90 percent on everything. I guarantee you right now, we work on 90 percent of all use cases. 90 to 95. Would you agree? Yeah, I think that a lot of people, there's a, there's a way to accomplish your use case.

Yeah. Although it may not be obvious is the one situation that it's like we're, however, this percentage is very, very easy. This percentage, you need to try the second or third methods to get it working. Yeah. So, oh, also, uh, he then mentioned that his biggest problem is testing. The iteration is really slow.

So he can't see what's happening in cloud in real time. Uh, I think that's just a by product at the moment. Right. Um, yeah, it's kind of difficult to do some things there. We can explore it. Yeah, yeah, yeah. Yeah. Right. So, um, the funny thing is, is here was the exact thing is it was like a payload issue for our cloud runner on our side where the automation was such a large scroll down step.

It had intertext 500 kilobytes. There was a maximum we allowed was going to be allowed to be stored in the database for the task history Oh, yeah, because if it exceeded that it wouldn't add the document adding that document is what shows it a new task And this was the first time like this. This is what's funny So I would call this a one in a hundred right, which is one percent, but it's not I thought about this It's technically one in 12, 000.

This is the first time we have had someone do this thing. Right.

So you think that a problem with it, reporting to their task history? Yeah, yeah, yeah. So if we have we have like around 15, 000 issues or something like that, like we have over 4000 paying customers and then they invite people. So let's just call. Let's just take the paid users. Right. Forty five hundred.

Okay, that is a one in 4, 500 thing out of all the things that have ever been reported this one edge case. Now, granted, it gets taken as cloud does not run. All right, because it failed on a certain thing. Now, we are here because within a day. We found out what it was, investigated that specific issue, and then within a day or two, we'll turn it around for people.

So we, we, we, our word to you is we're going to move really fast to get these edge cases. And then now this edge case is fixed for everyone. Correct, Kyle?

Yes, it is. Okay. Fine. All right. So this is, this is the reason why we're launching the community in the next month or so after I put all the content in there. It's just a lot of stuff. So, um, and I'll automate it. This is the reason why we do what we do. This is the reason why we allow this to be public because if we're going to handle every single edge case, right, we will handle them.

You just need to tell us, hey, so it's not, you can report it as clouds not running, but when in fact cloud isn't running for this specific thing, I wanted to be very clear on that. So, and like it was reported in bugs. So it's not a bug. Right? At all. We just haven't accounted for that edge case. That 1 in 4, 000 edge case.

Right? Like, that's what that is. We've never gotten that reported before, specifically. So then, it takes a day or two, we respond to it. That is how we need to go about this. It's very community driven. You need to tell us the esoteric edge case that you're working on. Within a couple days, we'll have the fix.

And then you'll be fine. Um, like they said, like, uh, he reported it, it would, it will be good. Um, so like, yeah. Is there any, is there any other example between like a bug and an edge case? I believe at this point, our stuff works 95 percent of the time. 95 percent of sites and cases. What do you think, Kyle? I think so.

I think it's just about exposing to everybody how they're going to do it, or the workarounds for when those, because a lot of sites are doing things that make it difficult for you to do what you want to do. Yeah. So sometimes when one method that works on one site doesn't work on another. So if you're, if you're limited, you're, let's call it your toolkit of building automations is limited to one method.

You can build on less than someone who has used other methods on other sites. So it just goes into what the course is talking about those things of different ways to accomplish things on sites. So when it's difficult, you can still do it. Right. So what he's talking about is we have we are launching a course Um, and that I'm putting together right now.

So where we have like our success stories and a lot of the, we're filming dozens and dozens of videos on this that will transcribe and make into content for this, for our school course, um, for you guys that you can like get started and say, Hey, my thing doesn't work on this piece. It's like, we will get to every single edge case.

Like, cause again, our goal is to boil the ocean and we're going to do it. So just have exercise some patience with us. Um, and by the way, Hans was amazing and he reported that it was working a day later, right? So we just have to work through those edge cases at this point In our journey, if something does not work, there are still bugs.

We still do have bugs, but it's either local compute, like your computer is very slow or very, it's not, it doesn't have like proper resources allocated to it. Or, like so you'll get like network issues or things like that. Or, it's just an edge case we haven't accounted for. I do believe we're out of the woods of critical errors.

Would you agree? We're out of the major things. I think a lot of people like see small issues with recording on sites. And they don't realize that like, on any note, I would say that, I would say that every no code RPA tool fails to handle iframes. And we're failing in handling like, iframes inside of an iframe that loaded really slow.

And that's the kind of situation we're at. The, the alternative is, is you digging into the website and you finding your own selector. Which obviously is not what we talk about doing at all. But that is a possibility. It is, you're able to do it. We just, we do good in almost all the scenarios where you shouldn't have to do that.

Yeah, great. And we'll make a video on that as well. So, okay, anything else to add? Again, so just as a reminder, if you run into an issue, just let us know. Most of the things, and a lot of people can attest to this, we usually get to and address edge cases within a couple days. And like, well, cause we're just curious about it.

So, and we're curious to help and like drive the community forward. So we appreciate Hans and Isaac and all these other people that throw stuff in here and let us know what edge cases they are and we'll investigate. Okay. Thank you so much to you guys. Um, you are terrific. Everyone out there, uh, and subscribe to the channel.

Say hi. Say hi. Okay. Talk to you later.