I've been observing a significant pattern in the technical discussions across the platform, particularly in the Infrastructure and Databases subforums. A substantial portion of threads are not about solving problems with an existing, chosen technology, but rather about the preliminary, comparative analysis phase *before* a decision is made. These threads often follow a similar structure: "We are evaluating X versus Y for our new event streaming layer," or "Need to choose a managed Postgres service, here are our constraints."
The current forum structure forces these discussions into subforums dedicated to a specific technology (e.g., "Databases"), which can lead to bias, fragmented information, and a lack of visibility for those specifically in a research phase. The "Site Feedback & Meta" is the only place to discuss this, but it's not the right venue for the technical content itself.
Therefore, I propose the creation of a dedicated **"Pre-Procurement & Technology Evaluation"** subforum. Its purpose would be to host structured, analytical discussions about selecting tools, services, and platforms. This would provide a neutral, comparison-focused space separate from implementation-specific support.
The scope would be deliberately broad but technically focused, covering:
* Head-to-head benchmark analysis and methodology (e.g., "Designing a fair load test for DocumentDB vs. Cosmos DB").
* Total Cost of Ownership (TCO) modeling for competing services, including egress, compute, and operational overhead.
* Evaluation criteria frameworks for specific domains (e.g., "What metrics matter most when choosing a CI/CD platform for a polyglot microservices architecture?").
* Requests for Information (RFI) styled posts where members detail their technical and business constraints for community input.
A key benefit would be the consolidation of knowledge. Instead of a fragmented discussion about AWS RDS vs. Google Cloud SQL in one thread and Aurora vs. Cloud Spanner in another, all these comparative discussions would be in a single, searchable location. This would create a valuable archive for future researchers. It would also clean up the technology-specific subforums, allowing them to focus more deeply on implementation, troubleshooting, and advanced usage of *already-chosen* tools.
To maintain quality, we might consider a posting template or strong guidelines to prevent vague "what's the best X" questions. A good post in this proposed subforum should mandate the inclusion of:
* Clear technical and non-technical constraints (latency requirements, team skill set, budget ceilings).
* Current architecture context.
* Shortlisted options already identified.
* Specific, comparable attributes under evaluation.
For example, a high-quality thread starter might include a structured comparison matrix in a code block:
```markdown
| Evaluation Criteria | Option A: Service X | Option B: Service Y | Our Weighting |
|--------------------|---------------------|---------------------|---------------|
| P99 Read Latency | <15ms | <10ms | High (40%) |
| Cost per 10k ops | $0.25 | $0.32 | Medium (30%) |
| HA Failover SLA | 99.95% | 99.99% | Medium (20%) |
| Managed Backups | 35-day retention | 21-day retention | Low (10%) |
```
This structured approach elevates the discussion from opinion to analysis.
I am interested in whether others find this gap in our forum taxonomy as pronounced as I do, and what potential pitfalls or scope creep we should consider for such a subforum.
Data over dogma
This is such a great observation. I see this constantly in the Product Analytics subforum too. Someone will post "Amplitude vs. Mixpanel for a startup with X needs," and the discussion inevitably gets siloed into Amplitude-specific advice, even when the core question is really about the *process* of evaluating and selecting a tool based on first principles.
Your point about bias is spot on. Posting a "Pre-procurement" question in a dedicated tech subforum primes the community to think from the perspective of that technology's users, who often (understandably) advocate for their choice. A neutral space would encourage framing the question around requirements, team skills, and trade-offs first, before we all jump in with our favorite tool.
My one caveat, and maybe the next thing to think about, is that pure comparison threads can become "bake-offs" without enough real-world context from the OP. A new subforum would need strong moderation to enforce that posters include their actual constraints, scale, and what they've already researched. Otherwise we'll just get "Which is better?" posts, which aren't useful for anyone.
I completely support this proposal. Your point about bias is crucial, and I've seen it play out negatively in vendor risk contexts. When someone posts "Which SOC 2 compliant CI/CD platform?" in a Jenkins-specific forum, the feedback skews heavily toward technical features, often overlooking critical procurement factors like contract liability clauses, data processing addendums, and the actual audit scope of the vendor's compliance reports.
A dedicated subforum could enforce a more structured evaluation template. Threads might require posters to outline not just technical specs, but also their compliance requirements, intended contract tenure, and risk tolerance. This would elevate the discussion beyond "which tool is better" to "which tool is a governable, contractually sound choice for our specific context."
My only addition is that such a space would need strong moderation to keep it from becoming a marketplace for vendor sales teams. The focus must remain on the methodology of selection, not just the outcomes.
RTFM — then ask for the audit
I agree this structural bias is a real problem, especially for performance-critical decisions. When someone asks "Kafka vs. Pulsar vs. Redpanda" in a streaming-specific subforum, the latency and throughput trade-offs get discussed in a vacuum, often missing the crucial pre-procurement context of team size and operational maturity. A junior team shouldn't be optimizing for the last 100 microseconds if they lack the expertise to operate the chosen system reliably.
A neutral evaluation space could enforce the inclusion of actual load-testing requirements in the opening post. Too many threads devolve into anecdotal performance claims because the poster didn't provide a concrete RPS target, p99 latency budget, or expected data volume. If the subforum template mandated this, we could skip the generic "X is faster" debates and focus on which tool meets *quantifiable* criteria under constraint.
My concern is dilution. Would this new forum simply become a graveyard for "vs." posts that get ignored because the specialists who know the technologies deeply don't monitor it? The technical subforums have concentrated expertise. We'd need a cross-posting or tagging mechanism to pull relevant experts into the neutral evaluation threads without letting the discussion revert to a home-team bias.
--perf