1478
Views
13
Comments
Solved
what difference between service action and server action

what difference between service action and server action

2021-09-06 15-09-53
Dorine Boudry
 
MVP
Solution

Hi @Andrew,

this is a much asked and answered question, so googling should give you many results.

See for example this documentation for a comparison.

@Manish, 

I'm not really convinced about your answer.

  • for starters, I don't think it is asynchronous at all, it runs in a separate process, and as a consequence has a separate db transaction, which makes it tougher to deal with in case changes have to be rolled back, but it does run synchronous.  
  • but also, I don't think their use case is specifically or primarily to integrate with external systems, it is about how your Outsystems producers and consumers are coupled.  
  • long running processes will still be long running, doing them in service actions won't help, only timers or (light) bpt can help with that

I think the most important reason to use them despite the added complexity, and despite maybe some performance overhead of extra transactions, is as an alternative to exposing server actions from a producer, to offer a loose coupling allowing more modular deployments of changes.

Dorine

2024-07-05 14-16-55
Daniël Kuhlmann
 
MVP

You are 100% correct,

I marked your reply as solution and unmarked the reply from Manish.

UserImage.jpg
Lukas Tomek

Seriously, the mods of OutSystems community should warn people posting AI responses and if they do not stop, they should be simply banned. I am new to OutSystems and I have to go throw wrong answers on the forums already several times? The AI answers should be also immediately deleted.

2024-07-05 14-16-55
Daniël Kuhlmann
 
MVP

If there is proof that replies are generated by AI, moderators mark them as AI and a such they are visible. And also not all AI generated replies are incorrect. But I don't think this thread has AI generated replies.

All that said, if you think a reply is incorrect, downvote it, or reply on it.

We cannot just simple ban people for posting an AI generated reply, for many reasons. For example, if the community member works for an outsystems customer, e would not be able to use his personal environment for example, nor access to the community pages that require a login, while he or she works for a paying customer.

All I all I think the AI generated content isn't a big issue as you think on the OutSystems community

UserImage.jpg
Lukas Tomek

Hi, thank you.

1) This thread has an AI reply that is already marked as a an AI reply.

2) These are your own words about the reply visible in this thread: "Sorry but there are several mistakes in this reply regarding service actions and it should not be marked as solution." (my take is it should be deleted, not just downvoted because it emerges in Google and also these times in AI answers)

3) Hopefully OutSystems can ban people in forums and let them continue to use all other parts of the platform. I understand it is not possible now but maybe it can be considered for the future. You can consider some people could want to grind points on the platform to look legit to new customers, so they would do it robotically to get high in leaderboards.

4) AI replies are a huge issue everywhere, on Reddit, Facebook, Twitter etc. AI comments can completely ruin a discussion platform. It is better to prevent it.

2024-07-05 14-16-55
Daniël Kuhlmann
 
MVP

alright I overlooked indeed that reply and my comments on it... sorry was in a hurry.

2021-09-06 15-09-53
Dorine Boudry
 
MVP

@Lukas Tomek ,

I am one of the few moderators that spend personal unpaid time on this, so I am going to share a few thoughts

On the division of labour

the moderator :

  • reads maybe 50 % - 70 % of ALL POSTS AND ANSWERS
  • checks if any question or answer is SPAM (most of that is done by OS spam filter now, only a handfull a month go undetected)
  • checks if the question is asked clearly and is not a duplicate
  • checks if the answers are correct and clear
  • opens and tests OML's to check if the given solution is correct and up to best practices
  • engages with posters if needed
  • Added to that since roughly 1.5 years, tries to gauge if an answer is AI generated, and if so, marks it
  • engages with the other moderators occasionally to discuss particular post(er)s

the beginner :

  • Googles an OutSystems concept he is learning about
  • finds some OutSystems learning material about the subject
  • also finds some posts on the subject of fellow community members who were, just like him, struggling with this and tried to help each other

So now you are asking me to do even more, and engage with people who post AI generated content, trying to dissuade them or else ....  Instead, I would like to ask you to apply some caution when reading anything here, and develop your own radar for AI generated stuff (we already try to mark all of them as such)

On the future of the forum

I personally profoundly dislike AI generated content, but I try to treat it the same as the other posts, if they are wrong or unclear, I act on that. If they are correct and helpful, why not ? I still like to feel a real human is behind a comment, but that's just me.

Right now, the forum still works and AI content is not too destructive yet.  If in the future it gets worse, where 1 in 3 posts is AI generated garbage that only confuses things instead of giving real answers, or an equal threat : AI generated SPAM that gets through the spam filters, I can imagine I won't find the energy anymore to moderate that on a daily basis, and I can imagine people will resent coming here if they have to wade through all of that.

But that's going to be up to OutSystems themselves, to come up with an answer to what the forum will be in the age of AI.

Dorine

2021-05-18 02-27-17
Manish Gupta
Champion
AI Generated

Hey Andrew

In OutSystems, Service Actions and Server Actions are both mechanisms for implementing business logic within an application. Let me put some lights for your understanding - 

Service Actions:

  • Service Actions are primarily used for integrating with external systems or services, such as web services, databases, or APIs.
  • They allow you to define input parameters and output structures and encapsulate the logic required to interact with the external system.
  • Service Actions are typically asynchronous and can perform operations that involve network communication or long-running processes.

Server Actions:

  • Server Actions are used for implementing reusable blocks of server-side logic within an application.
  • They are typically used to implement complex business rules or calculations that must be shared across multiple screens or modules.
  • Server Actions are synchronous and execute within the server-side context of an application.


Example: Let's say you're developing an e-commerce application in OutSystems. You might have the following scenarios:

Service Action: You want to integrate your application with a payment gateway service. You would create a Service Action called "ProcessPayment" that takes input parameters such as the payment amount, customer information, and payment method. This Service Action would be exposed as a web service, allowing external systems to invoke it and process payments.

Server Action: In your e-commerce application, you have a checkout process that calculates the total order amount, applies discounts, and updates inventory levels. You can create a Server Action called "CalculateOrderTotal" that encapsulates this logic. It would be used internally within your application and called from the checkout screen or other server actions involved in the ordering process.


This answer was AI-generated. Please read it carefully and use the forums for clarifications
2023-06-21 08-58-47
Andrew Mahfouz 

Thanks 

2024-07-05 14-16-55
Daniël Kuhlmann
 
MVP

Sorry but there are several mistakes in this reply regarding service actions and it should not be marked as solution. 

Just one example that is 100% incorrect:

This Service Action would be exposed as a web service, allowing external systems to invoke it and process payments.

This reply reads as if it was generated with chatGPT. Something we recently see more on the forum and ideas pages, unfortunately chatGPT doesn't generate 100% correct answers (yet).

2021-05-18 02-27-17
Manish Gupta
Champion

Indeed a valid observation. Thanks for the corrections @Daniel Kuhlmann & @Dorine Boudry 

PS: I did 3-4 Pilot posts from GPT yesterday. We are waiting for the community's reaction to others as well. We're doing case studies on People's Reactions, especially developers' community (Budding to Experienced), toward it. 

2021-09-06 15-09-53
Dorine Boudry
 
MVP

Ok,

so here is some honest and heartfelt reaction from one member of the community :

I saw these GPT posts in the last days, they were somewhere between low quality and blatently wrong, i did not have the energy to react to all of them.

My reaction to your experiment is that these are not the terms under which I am willing to participate on the forum, I am here to help others and learn in the process, I did not volunteer to take part in case studies as a lab rat.

In my opinion maybe it is time for the community guidelines to be adjusted, and extend the "Don’t Claim the Work of Others as Your Own" with some guidance on only using gpt after disclosing that you do.

Not doing studies on people without their prior consent falls under "Be Considerate" and "Be Respectful" that are already in there.

Dorine

2024-07-05 14-16-55
Daniël Kuhlmann
 
MVP

Fully argree with you @Dorine Boudry asking the OutSystems community team to look at this and also other community members that just copy/paste incorrect chatGPT replies. 

This is a trend that is reducing the quality of this community. It is really a pity that people are so naive to think they can blindly trust an AI generated reply, are too lazy to learn and think themselves or trying to play the gamification build into the community pages.

Over the years I worked on the OutSystems community, moderators have spotted a handful of people trying to play the systems to up their community ranking score in creative ways. But now with chatGPT it looks like more and more people think it is OK, to do so. IT'S NOT.


2021-09-06 15-09-53
Dorine Boudry
 
MVP
Solution

Hi @Andrew,

this is a much asked and answered question, so googling should give you many results.

See for example this documentation for a comparison.

@Manish, 

I'm not really convinced about your answer.

  • for starters, I don't think it is asynchronous at all, it runs in a separate process, and as a consequence has a separate db transaction, which makes it tougher to deal with in case changes have to be rolled back, but it does run synchronous.  
  • but also, I don't think their use case is specifically or primarily to integrate with external systems, it is about how your Outsystems producers and consumers are coupled.  
  • long running processes will still be long running, doing them in service actions won't help, only timers or (light) bpt can help with that

I think the most important reason to use them despite the added complexity, and despite maybe some performance overhead of extra transactions, is as an alternative to exposing server actions from a producer, to offer a loose coupling allowing more modular deployments of changes.

Dorine

2024-07-05 14-16-55
Daniël Kuhlmann
 
MVP

You are 100% correct,

I marked your reply as solution and unmarked the reply from Manish.

UserImage.jpg
Lukas Tomek

Seriously, the mods of OutSystems community should warn people posting AI responses and if they do not stop, they should be simply banned. I am new to OutSystems and I have to go throw wrong answers on the forums already several times? The AI answers should be also immediately deleted.

2024-07-05 14-16-55
Daniël Kuhlmann
 
MVP

If there is proof that replies are generated by AI, moderators mark them as AI and a such they are visible. And also not all AI generated replies are incorrect. But I don't think this thread has AI generated replies.

All that said, if you think a reply is incorrect, downvote it, or reply on it.

We cannot just simple ban people for posting an AI generated reply, for many reasons. For example, if the community member works for an outsystems customer, e would not be able to use his personal environment for example, nor access to the community pages that require a login, while he or she works for a paying customer.

All I all I think the AI generated content isn't a big issue as you think on the OutSystems community

UserImage.jpg
Lukas Tomek

Hi, thank you.

1) This thread has an AI reply that is already marked as a an AI reply.

2) These are your own words about the reply visible in this thread: "Sorry but there are several mistakes in this reply regarding service actions and it should not be marked as solution." (my take is it should be deleted, not just downvoted because it emerges in Google and also these times in AI answers)

3) Hopefully OutSystems can ban people in forums and let them continue to use all other parts of the platform. I understand it is not possible now but maybe it can be considered for the future. You can consider some people could want to grind points on the platform to look legit to new customers, so they would do it robotically to get high in leaderboards.

4) AI replies are a huge issue everywhere, on Reddit, Facebook, Twitter etc. AI comments can completely ruin a discussion platform. It is better to prevent it.

2024-07-05 14-16-55
Daniël Kuhlmann
 
MVP

alright I overlooked indeed that reply and my comments on it... sorry was in a hurry.

2021-09-06 15-09-53
Dorine Boudry
 
MVP

@Lukas Tomek ,

I am one of the few moderators that spend personal unpaid time on this, so I am going to share a few thoughts

On the division of labour

the moderator :

  • reads maybe 50 % - 70 % of ALL POSTS AND ANSWERS
  • checks if any question or answer is SPAM (most of that is done by OS spam filter now, only a handfull a month go undetected)
  • checks if the question is asked clearly and is not a duplicate
  • checks if the answers are correct and clear
  • opens and tests OML's to check if the given solution is correct and up to best practices
  • engages with posters if needed
  • Added to that since roughly 1.5 years, tries to gauge if an answer is AI generated, and if so, marks it
  • engages with the other moderators occasionally to discuss particular post(er)s

the beginner :

  • Googles an OutSystems concept he is learning about
  • finds some OutSystems learning material about the subject
  • also finds some posts on the subject of fellow community members who were, just like him, struggling with this and tried to help each other

So now you are asking me to do even more, and engage with people who post AI generated content, trying to dissuade them or else ....  Instead, I would like to ask you to apply some caution when reading anything here, and develop your own radar for AI generated stuff (we already try to mark all of them as such)

On the future of the forum

I personally profoundly dislike AI generated content, but I try to treat it the same as the other posts, if they are wrong or unclear, I act on that. If they are correct and helpful, why not ? I still like to feel a real human is behind a comment, but that's just me.

Right now, the forum still works and AI content is not too destructive yet.  If in the future it gets worse, where 1 in 3 posts is AI generated garbage that only confuses things instead of giving real answers, or an equal threat : AI generated SPAM that gets through the spam filters, I can imagine I won't find the energy anymore to moderate that on a daily basis, and I can imagine people will resent coming here if they have to wade through all of that.

But that's going to be up to OutSystems themselves, to come up with an answer to what the forum will be in the age of AI.

Dorine

Community GuidelinesBe kind and respectful, give credit to the original source of content, and search for duplicates before posting.