A Twitter feed app for restricted use blocked from write privileges


#1

There is a public archive called arxiv.org officially ran by Cornell University but used by scientists the whole world over for posting research papers for free access for everyone. The subject areas are physics, mathematics, computer science, statistics, quantitative finance and quantitative biology, each with many subareas.

My app posts to Twitter short summaries of the papers posted each day there, in various subcategories for mathematics. There are in total a bit more than 30 subcategories and for each I created an account. My application writes only to these IDs and there is no abuse.

I checked carefully that arXiv terms of service allows this. I also checked carefully that it does not violate Twitter terms of service, or at any rate falls within the possible exceptions for community benefit. I have written it up in detail at my URL:

https://www.notionsandnotes.org/pages/arxiv-twitter-feeds.html

As can be seen, I carefully verified each provision. This was a volunteer effort for community benefit without spamming anybody. I believe there should be more dissemination of public research free of cost. It may be of benefit to Twitter also to include such high quality content as daily feeds. It is an unexpected setback that the app got blocked.

Furthermore, bot/accounts for a similar purpose are running for years; for example, by @vela running accounts with IDs: mathACb mathAPb mathATb mathCAbot mathCObot mathCTbot mathCVb mathDGb mathDSb mathFAbot, etc., about the same number of accounts as mine. The difference with my app is that the previous one is restricted to 140 chars for the summary whereas mine compiles the summary into a one-paragraph image using LaTeX, the document typesetting system for mathematicians, and properly renders all mathematical symbols, etc., and I believe this is a significant improvement and will make a lot of difference for the users.

I hope the explanation satisfactory and my account write privileges are re-instated, without which tweeting is impossible. As I put a lot of effort into it for a project of research community benefit, it will be tragic if it all goes waste because you blocked it. As my effort is nothing like spam, it is disappointing to be treated like a spammer. Hoping to hear a resolution for the problem.


#2

Thanks for the explanation, but this is not the correct place to post it - you’ll need to respond to the emails you’ve been sent regarding the write-restriction, or raise a ticket via https://support.twitter.com/forms/platform - we are unable to take action or comment on individual applications or situations on these forums.

Note that most of the systems involved in first response to application behaviours like this are automated, and apply rules based on common automation bad practices and machine learning. The automation rules generally have more information on this, but I do understand how frustrating it is when you’re trying to build something like this, and it is difficult to document every case that would cause an app to trigger our antispam system, because it is an adaptive set of rules that has to respond to new attacks every day. If you’re interested in learning more about it (it is called “Botmaker”) then you can watch a video from the team from our Flight conference last year.


#3

Thanks for the reply. I submitted a ticket and my app was re-enabled. Thanks.

However, I also got the following:

Going forward, carefully ensure that your application adheres to Twitter’s policies:

Developer Agreement and Policy
Automation Rules Best Practices
Developer Display Requirements
Twitter Rules

Please note, applications found to be violating policy again are subject to permanent suspension.

I understand that spammers are very clever and it is a machine-learning algorithm but it is a mild threat to have an eternal Damocles’s sword of “permanent suspension” for unknown reasons. In any case, thanks for the fast response, both here and for re-enabling the app in the ticket.


#4

Noted, and thank you for that honest feedback. It is not our intention to come across in that way, but rather to be clear about how and why rules are enforced. I’ll see if we can do anything to change the language to feel less intimidating, but that is controlled by a different team to the one that looks after these forum.

One thing I’ve seen in the past is that regularly posting images has triggered some of the suspension rules, so it could be that this aspect of your bot, whilst very cool and potentially nice for users, might affect the behaviour of the system.

Thanks for building on the Twitter platform!


#5

The image explanation does make sense and is a possible answer. However I have also seen Twitter bots that automatically post highly rated images from reddit; so it seems it is allowed.

To feel less intimidating, firstly I have to agree that the approach given above may indeed be a right for spammers. But when one is not a spammer and a genuine case, a softer message may be considered.

Thanks for the kind words, and for taking time to explain and listen to feedback. Although I was initially a bit dismayed, by now you have made me feel good and appreciated, and thank you for the patience and attention!