Because comments are not Big Data.
Yep, you got it.
Also MIT licensed.
Please go ahead and kill Facebook comments.
I strongly dislike realname policies in this day and time where a single remark can cost you your job.
I'd like to understand the significance of this. I'm extremely interested in comments as big data (to try to understand social networking patterns better from studying the structure of nested comments) and I'd like to get a handle on the opposite point of view - what is it that you prefer about this approach?
While this might be interesting to you it is in the best interest of the site owners and commenters worldwide that not everything they write is instantly correlatable to anyone who happens to work for Facebook or be in a position where they can somehow buy or otherwise demand this data from them.
I guess collection, processing and stockpiling of user data should be regulated for the benefit of consumers, moderators and world peace.
I was confused by the original post because a) I'm not sure why using SQLite makes comments into 'not big data' (is it extraordinarily hard to get statistical information out of SQLite or something?) and b) even if a blog or website is completely private, might not the owner want to analyze what's going on within the private social network they've created?
Also, I question your blanket assertion about what is or in't in the interest of other people. You seem to have taken my prior comment as an intent to exploit comment data for financial gain.
Then you might get a license from me. And if you haven't tested with the HN dataset that is where I would start.
I was confused by the original post because a) I'm not sure why using SQLite makes comments into 'not big data' (is it extraordinarily hard to get statistical information out of SQLite or something?)
No, getting data out of SQLite is easy. But I have yet to hear anyone using SQLite for anything that would be called big data in a way that makes sense. (Yes, my previous employers file share where we could upload 5GB files doesn't qualify.)
I think the comment on the website is supposed to be funny in a bitter way.
and b) even if a blog or website is completely private, might not the owner want to analyze what's going on within the private social network they've created?
I have yet to manage to think of a private website that might contain enough non-spam comments to qualify as big data.
And for small and medium size data plain SQLite is wonderful : )
What I and others are skeptical to is REAL big data, where certain companies scoop up everything they can get their hands on. This is an often dirty, potentially really harmful raw material as we have seen in articles about the moderation teams that makes sure Facebook and others stays reasonably clean.
With a little bit of processing this "nuclear waste" is also possible to weaponize. Two ideas off the top of my head:
- Facebook and others can easily name thousands of people who waste their employers time during the day.
- or post a number of people who run pseudonymous accounts for various reasons but have failed to maintain proper opsec.
Mine uses Redis or SQLite for storage, and uses markdown. The biggest difference is mine is flat. I keep thinking I should rework it to allow nested comments, and avatars, but at the same time it is nice and simple as-is..
Demo - https://tweaked.io/guide/demo/
Can it though? Because I see some vile comments made by people who appear to be using their real name, photo, location information on their profile.
Big data is mostly used when you want to monetize those comments.
When I first set up isso, I had a hell of a time getting it configured and running, following the pretty poor official documentation. So I visited the Github repo to ask for help but, after seeing the dev's condescending attitude to other people in difficulties, I thought better of it and [fortunately] managed to muddle through by myself.
So you tell them to go use Disqus if they need Gravator since they're basically stating they don't care about anti-tracking.
You then also state that someone else should do the work of integrating with Libravator.
It reads a bit gruff and asshole-ish.
On the other hand, TANSTAAFL and either someone should pay you or do the work to do the Libravator integration, and it does seem very odd to pick up Isso and then want to use an avatar service that also does tracking.
Presumably those people hate Disqus for other reasons, but it certainly doesn't have to be your concern.
I'd probably write the same kind of tone in a FAQ entry as well, because if its being given away for free then the flip side is that users on the internet shouldn't be demanding assholes and want things for free that you have zero interest in providing for free, and I think you make your position reasonably clear.
Isso looks nice, especially like the code blocks and backtick support. However, social login is a must have. The ability for anybody to put a name and email is bound to be a spammers delight.
Why should I use Isso over Disqus?
They've had the program in place for years (~2012) and it is opt-in. Author didn't seem to contact Disqus to see if this was a mistake or intentional, and doesn't mention if the domain had the Revenue option enabled. He just says 'I remember once clicking an unrelated option about showing inline ads, that should have been good enough'.
Seems like a senseless ragequit.
It is that simple. :)
Your server, your data.
Disqus is offering you a free service for a reason: they want to make good use of the data that your site comments contains.
They also want to hi-jack the community that you have built around your site - or rather: what Disqus has fooled you into believing that you have.
It is not a product.
And, as such: doesn't need to market itself.
Use it if you want to. Or not.
I wanted to use it, but couldn't make it run - I am not a Python guy, so I went for HashOver (PHP) instead.
Not because HashOver is better, it's not, but because I need to be able to modify/fix the software, and I know PHP better than Python.
But I still believe that Isso is the best self-hosted and open source Disqus alternative in existence.
I looked into using Discourse to do something like this, but they refuse to reduce their memory usage (2GB minimum required, I think)
coming from a Python project...
The documentation doesn't mention much there, so I'm afraid I'm stuck. Has anyone managed to set this up? I'll open an issue with them to improve the docs.
Regardless, it works quite well now, thank you!