What Should OpenSSL Improve for Individual Contributors?
Hello everyone, hope you are doing well!
During the last OpenSSL Corporation Face-to-Face meeting, I had the opportunity to present feedback collected from the Individuals Community. The presentation was based on Individuals community discussions, survey responses, conversations with individuals at various conferences, and general feedback received over the past several months.
The presentation highlighted what the community appreciates, as well as several recurring pain points(slide 04 and 05). I'd love to hear from an even wider audience.
Do these pain points resonate with your experience?
Do these pain problems(from slide 04 & 05) resonate with your experience? Vote below and let us know.
poll by Aditya Koranga Closed Wed 15 Jul 2026 3:00PM
Results
| Results | Option | % of points | Voters | |||
|---|---|---|---|---|---|---|
| Yes, I relate to these pain points. | 50 | 5 | ||||
| No, my experience is different.(pls tell ur pain points) | 50 | 5 | ||||
| Not voted | 145 |
For me the review process is one of the pain point
When I looked at the codebase for the first time, I found it difficult to understand due to the extensive use of macros
Yeah, I agree with the point (especially page 5).
There are already so many open PRs and issues that it can discourage new contributors. I sometimes find myself wondering whether it's really worth spending time preparing another PR or opening another issue if it's just going to sit in the backlog indefinitely.
It starts to feel like adding another drop to an already overflowing bucket.
A better approach must be taken as soon as possible.
While these are all valid problems, they are well-known and acknowledged; for some, even solutions have been discussed in Brno (e.g., PR review and issue resolution process). Thus, I'd dare state that a bigger pain point is a lack of communication in (and/or participation towards) moving solutions forward: Just another poll with ultra-low participation feels as irrelevant as proposing things to which no feedback is given (https://openssl-communities.org/d/96qzarIB/topics-for-brno-f2f-meeting).
I understand some of the points, but others are unclear. It's also not clear what should happen as result of this vote.
@Kurt Roeckx Could you let me know which points you found unclear?
The goal of this poll is to identify the most common pain points, bring them to the community's attention so that they can be prioritised and we can try to find some solutions around them, and also uncover issues the project may not yet be aware of.
PR feedback loop needs to be improved: What are the problems seen with the feedback loop? It takes too long? It's unclear what needs to change, or something needs to change at all? There are conflicting feedbacks? The feedback doesn't cover the whole PR, and you need to wait for someone else? Or something else?
Making the contribution process easier for Individuals: What does this mean? What are the issue they see? This again seems something that doesn't actually identify any issue, just that there might be some unidentified issue.
"can AI be utilised to reduce these numbers" is not an issue, it might be a solution.
-
"Looking back, how have we improved after receiving several criticism. Did
we make any changes/progress?". This is again not a specific problem.
So maybe this is all just formulated in a strange way, and you're just looking for feedback, and that are some of the topic you think about people might want to comment on.
@Kurt Roeckx oh ok i got it now. yes i think some of the statements could have been formulated in a better way and maybe the actual prsentation explaining these slides would have made it easier to understand. Btw these pain points/feedbacks are collected from the individuals through several engagements. and I wanted to show the overall summary of them alongwith other things.
We have observed that pull requests (PRs) often go unnoticed or do not receive timely, constructive feedback. This delays the review and merge process, and can discourage contributors from submitting future PRs. this one
Given the large number of open issues and PRs, contributors often find it challenging to identify a suitable issue or feature to work on based on their experience and capacity, determine the appropriate person to contact for guidance, understand whether an issue is already being worked on, or whether an existing PR is likely to resolve it before they begin their own work, etc.
Yes The question, "Can AI be utilized to reduce these numbers?", was intended as a proposed solution to the pain problem of "large number or issues and open PRs". It was included to encourage discussion and explore possible AI-assisted approaches during the Brno F2F meeting.
The last one can been seen here for refernce.
Build/test cycle times on x86 NonStop
@randallbecker
Heh, OpenVMS was a similar pain point. But it's admittedly quite slow, even on x86_64.
Putting on my individual contributor hat:
- The documentation a huge pain point.
- I'm concerned about the criticism OpenSSL has received recently.
- The backlog and review process are daunting if you have anything complicated to submit. I'd love to get more reviewers and more automated tests to help.
@Jon Ericson Completely agree with/support your items 1&3. Just with item 2 I wonder whether you could be more concrete which criticism you mean/are mostly concerned about (to make this actionable)?
@baentsch The space for a vote comment was too short to include details, but I'm specifically thinking of these articles. I am working on some blog post responses that highlight the concrete things the project has actually done in the last year or so.
I agree with others who have said that these are valid issues, but that they are either well known, or expected with a mature industrial codebase. At my place of work, we've been focussed on PQC algorithms, so if anything, I'd be curious if there is an OpenSSL organization-wide roadmap or plan published regarding post-quantum cryptography. Please excuse me, though, if it there has been; I've not gone looking for it very hard.
@dan pittman I agree these are well-known issues, but we still receive this feedback today. That suggests they've either not been prioritized, haven't received enough attention, or a suitable solution hasn't been found yet.
Yes, people have asked us whether openssl project has any PQC roadmap or not and they would be interested to see a roadmap like that(this point is also mentioned in the slides btw)
I second @Nikolas Gauder. The codebase is overwhelming for a beginner.
@S E Hmm -- I'd argue that all but the most trivial of code bases are overwhelming for beginners (unless they're geniuses or AIs :-). What make OpenSSL dauting is its long history and various architecture decisions that are not documented anywhere (echoing @Jon Ericson s comment above): Maybe time for an effort to get ADRs documented?
Item removed
I missed the poll deadline but I'd say:
The PR backlog is a problem
The codebase might be difficult for "beginners" but do we want PRs from contributors (or AIs) that don't understand it? "Solving" this by dumbing down the code seems like a distraction.
(that being said, the previous comment about macro overuse is valid)
@Simon Cornish Regarding the second point, every contributor starts somewhere so It would be great to have more beginners join, learn about the project, and contribute. If the project has a steep learning curve, good documentation plays a crucial role, and as mentioned above, documentation is one of the biggest pain points.
Middle of July voting... heh
I can relate to the codebase difficulty pain point. The codebase is humongous, and includes a pile of churn that's pretty daunting (just see everything that make update does). We also have a tendency to pile stuff into existing libraries and files where they don't really belong, which makes it hard to find what you want to work with.
@Michael Baentsch yes, proper communicaiton is required and thanks for stating(this is also mentioned in the slide no. 8 btw). We will do our best to improve in this area- yes One obvious straightforward way is that BAC/TAC needs to be more active and engage with the comments/feedbacks provided