Quora or Stackoverflow kind of application

Need help
1

Is to possible to create an Application like Quora or Stackoverflow using bubble?
Where user can ask Question and will receive answers on it.

2

Yes, absolutely. Search the forum for questions on building messaging features since that would be the core type of functionality you’ll need. Your data structure would include something like this –

Question

  • Title (text)
  • List of Answers (list of Answers)

Answer

  • Text (text)

You’d layout your page to display a list of Questions, which can be searchable, where each question can display its list of answers. Workflows will allow you to create a question as well as create an answer and add that answer to a questions list of answers. This is super high-level, but play around with it and search the forum for examples!

Gaby | Coaching Bubble

1 Like
3

This is definitely possible, keeping in mind there are millions of possible ways to do this depending on your needs/wants/wishes. Have you followed through and completed the tutorial lessons? That will give you a better understanding of what Bubble is capable of. There is also this page: https://bubble.io/documentation

I’d say expect to build more than one version of your app over time. Start simple and expand out once your core concept is proven. The forums are full of helpful conversation threads, so when you have a specific question try doing a search here first. For example there are quite a lot of “Can I build xyz type app?” topics that have been covered in detail.

Let me know if you have specific questions and I can help out where I can :slight_smile:

1 Like
4

@romanmg Thanks for your time and reply.
Yes, this is super high level and do want to give it a try.
when will video course come in coaching bubble?

5

@philip Thanks for your time and reply.
I searched through the discussion forum and I haven’t found anything related to my Question that’s why I asked can you give me brief Idea how I can proceed in order to build this application?

6

The interactive walkthrough tutorials are going to show you the ropes of basic Bubble functionality, so definitely start there. I’d suggest not building your end-goal app as your very first Bubble app. Instead, build out a few of the example tutorial apps and get a hang of what all the different elements do. Once you understand how the elements work (even a vague idea), you should be able to reverse engineer the features you want in your app and break that down into bite-sized pieces composed of elements and workflows.

As @romanmg pointed out, [quote=“romanmg, post:2, topic:15028”]
Question

  • Title (text)
  • List of Answers (list of Answers)

Answer

  • Text (text)
    [/quote]

(text) = Bubble text element
(list of xyz) = Bubble repeating group element containing a specific data type (either Question or Answer)

Once you get to the point where you’re placing elements on the screen and building basic workflows, posting in the forums with screenshots and specific use-case scenarios and problems you’re having will allow us to give specific feedback and recommendations. I don’t want to 1) limit your thinking by giving you a very bare bones example or 2) go above and beyond what you’re trying to do and confusing you in the process.

2 Likes
7

@philip Thanks for your time I got it I will start working on it.

8

I think it’s important to highlight that Bubble may work well to build an MVP for this, but it has some big limitations that limit what it can do for a production grade application in this space. For example:

Quora uses a sophisticated search capability that can search for multiple words within a title, description, topic tags, answer fields and will also look for synonyms, will use sounds like logic, will use natural language processing, etc. They rely on this search capability for most of user’s navigation on the site, and without it Quora wouldn’t be nearly as useful because users couldn’t find the good content that exists. Whereas, Bubble only supports full word match (i.e., if you type in ‘car’ it won’t find the word ‘cars’ in the results because the full word doesn’t match perfectly. I believe it must be the same case too so searching ‘car’ won’t find the word ‘Car’ in the results. There are workarounds that allow it to be case insensitive and do partial word matches in Bubble, but these are harder to implement and do not show results to users near as quickly as I’m sure you’d like it to. I haven’t heard of anyone finding a way to do multi-word searches in Bubble or use synonyms, etc. So, it’ll be much harder for people to use search to find the results they’re looking for.

This is just one example. So, yes, Bubble can help you create something that looks like Quora or Stackoverflow. However, it will have serious limitations as well.

1 Like
9

@sridharan.s ya I understand we can implement something that looks like Quora or StackOverflow but with serious limitations as well but is there any way to improve matching algorithm by integrating some NLP in the application.

10

There is a Blockspring API integration (paid, starting $10/mo) that may have some blocks up for the task at a MVP level: https://open.blockspring.com/browse%23Text%20Analysis%20%26%20Machine%20Learning

I haven’t tested these myself. I imagine you would need to have multiple data stores outside of Bubble that are responsible for the NLP-oriented tasks.

11

There are ways to integrate Bubble with 3rd party services or create your own algorithm and have Bubble interact with it via an API (e.g., using Blockspring). So, this type of solution can be really robust as it allows custom code. However, getting data back into Bubble via their API is really slow. So, it works well enough if you’re loading 1 result (i.e., 1 row) in real-time for the user or if you’re able to pre-load results in the DB for a user before they need them (for example, if you calculated the order of each person’s personalized feed each day using the API and then stored that in the database ahead of time in case they come to the site that day).

However, if you’re looking to show the user a list of, say, 10 search results when they type something in then that may take 5-10 seconds just for data to get back in from the API (not including 1) send time, 2) calculation time, or 3) bubble displaying it from the DB to the UI). Those first 2 steps can be really fast (e.g., ~1 second) and the 3rd step varies, but for now getting data back in is slow.

So, perhaps you can be creative and find solutions for most things. However, offering users a great search experience that works well and quickly doesn’t seem viable today. Perhaps there’s a solution that I’m missing though, and Bubble definitely continues to build out new capabilities. I hope they’re focusing more and more on building capabilities for enterprise clients with production grade apps, but I haven’t heard from them one way or another.

1 Like
12

Thanks for stepping in to drop some knowledge on us @sridharan.s, always appreciated! Great points you bring up.

13

Thanks for dropping knowledge @philip and @sridharan.s.
Bubble community is Quite helpful in comparison to that of other CMS community.
and @sridharan.s as mentioned by you even fetching 10 results from the API will involve 3 steps and it will take some time what if we need to fetch huge data then calling an API will not be an optimized solution.I think this functionality of integrating with NLP can be done by creators of bubble itself.

14

Yeah, I understand the sentiment as I often tell myself the same. However, there are lots of things for them to do before they’d build NLP into the search – so much so that they probably won’t ever get to it. Even with just search, it’d be nice if they enabled 1) case insensitive, 2) fuzzy search, 3) phrases, 4) elastic search across database tables, etc. Execution wise, they may build a more robust version on current search architecture first as a major feature release, then perhaps they’ll integrate Solr or Estaticsearch for a future major feature release which would get us the core search capabilities we’re using to seeing on big websites. After that, then perhaps there’d be a way for them to add NLP, but NLP doesn’t work all that well for generalized topics and instead generally requires a lot of extra effort to train the system on the language usage for your specific content domain. And, I can’t imagine this is something that Bubble would ever choose to take on, and definitely not within the next handful of years.

Besides, we’re missing a lot of other capabilities that people need for production applications. NLP is, at this point, way beyond the scope of what’s practical. Search that is usable by end users is within reach and is necessary for practically all apps these days. Let’s hope they start with that, along with the other core capabilities required for production grade applications that are still missing.

2 Likes
15

This topic was automatically closed after 70 days. New replies are no longer allowed.