Summary

  • ARIN's public 2021 election-forum transcript records Karl Morin describing work with smaller networks on IPv6, BGP, and peering while serving on the QIX board. CANIX now lists him as chair, Corporations Canada lists him as a director of the active CANIX entity, and the Global Peering Forum independently attributes a current presentation about Internet scale, compliance, and service needs to him.
  • Morin's signed CANIX note presents consolidation, a 24/7 network operations centre, 400G-capable sites, Layer 3 functions, and multiservice ports as answers to scale and continuity constraints. Those are attributed operating claims rather than independently measured outcomes. The useful test is whether the resulting platform keeps interconnection records, authority, costs, and failure responsibility legible to the networks that depend on it.

A Person-Level Record Built Around Operations

Karl Morin does not enter the public record used here through a generic executive biography. The strongest person-level source is an official transcript from the American Registry for Internet Numbers. During ARIN's 2021 election forum, Morin introduced himself as president of Hive Data Center and a board member of the Montreal Internet Exchange, QIX. He described working across peering and technical deployments and helping smaller networks move through IPv6 deployment and Internet operational processes.

The transcript matters because it records a concrete operating problem in Morin's own words. Smaller organizations may need BGP because their services, customers, or resilience requirements have outgrown a single-provider design, yet they may not have a large engineering team or a policy specialist. The difficulty is not only typing a router configuration. It includes obtaining resources, understanding routing relationships, coordinating with other networks, managing risk, and finding time to participate in the institutions that shape those processes.

Morin's remarks did not present a measured case study. They did not identify the networks he assisted, quantify the deployments, or document their outcomes. The transcript should therefore be used as evidence of his public operating focus, not as independent proof that every intervention succeeded. That boundary is important because a public candidate statement is still a self-account, even when it appears in an official forum.

The later records establish continuity without erasing the boundary. CANIX's official site lists Karl Morin as chair. Corporations Canada lists him as a director of CANADIAN INTERNET EXCHANGES CANIX INC. The Global Peering Forum's 2026 agenda identifies him as the CANIX speaker for a session on adapting to Internet scale, compliance, and service needs. These sources show that his public work remained attached to Internet exchange governance and operating questions.

The combination is stronger than a contact-only registry row or an event-attendance list. ARIN captures Morin explaining a person-level practice around BGP, IPv6, peering, and smaller networks. CANIX records his current governance role. The federal registry supplies a separate legal-entity record. GPF attributes a specific operator-facing topic to him.

None of those records turns Morin into the sole author of QIX or CANIX. Internet exchanges are shared systems involving members, directors, staff, vendors, data centres, route servers, transport providers, and other operators. The public evidence supports a profile of one named decision entity, not a hero narrative.

That distinction shapes the article. The relevant question is not whether Morin can be celebrated as a visionary. It is whether the operating model he describes makes the exchange more accountable and usable at the layer where networks actually connect.

The Small-Network BGP Problem

BGP is the protocol through which independently administered networks exchange reachability information. In practice, operating BGP requires more than owning a capable router. A network needs an autonomous-system identity or an appropriate relationship with a provider, address resources or delegated space, filtering rules, routing policy, monitoring, contact records, and people who understand how changes can affect traffic beyond the organization's boundary.

Large carriers and content platforms can distribute those responsibilities across specialized teams. A smaller provider, hosting company, enterprise, or regional operator may have only a few people covering systems, security, customer support, procurement, and network engineering. For such an organization, every additional coordination surface has a cost.

Morin's ARIN remarks focused on this asymmetry. He described smaller businesses that did not have the time or resources to participate deeply in policy processes, yet still needed practical paths through IPv6 and BGP deployment. That account identifies a real class of constraint without proving how common it is or how well any particular solution works.

An Internet exchange can reduce one part of the coordination burden. Networks connected to a common switching platform can establish peering relationships without purchasing a separate physical circuit for every bilateral relationship. Route servers can simplify session management where members choose to use them. Shared technical documentation and responsive support can lower the cost of joining and maintaining the service.

The exchange cannot remove operator responsibility. A network still has to decide what routes it will announce and accept. It has to validate filters, manage redundancy, monitor sessions, and understand the consequences of a mistake. An exchange that makes connection easy but leaves policy opaque may merely relocate complexity.

This is why the quality of an IXP should not be reduced to a membership count or a promotional claim about keeping traffic local. The operational questions are more specific. Are the peering LAN and route-server records accurate? Are changes communicated? Can members understand who is responsible for each service? Is support available when the failure crosses organizational boundaries? Can a member leave or change its relationship without losing control of its routing identity?

Morin's earlier public focus on small networks gives those questions a person-level anchor. His account implies that access is not simply a port price. It includes the human and institutional work required to turn a connection into a maintainable routing relationship.

The current CANIX platform should therefore be assessed against that problem. A national or multisite exchange can aggregate expertise and standardize support. It can also create a larger control surface. The test is whether the organization reduces unnecessary friction while keeping technical authority and responsibility visible.

What QIX Represented in the Earlier Record

Morin's 2021 ARIN statement connected him to QIX, the Montreal Internet Exchange. The later CANIX note gives QIX a longer institutional history. It says the exchange began within the Quebec research and education network environment in the 1990s and later became an independent nonprofit organization focused on public interconnection in Montreal.

That history is presented by CANIX and should be read as the organization's account. The federal corporate record independently confirms a narrower governance fact: the Montreal Internet Exchange corporation became inactive after amalgamation into the current CANIX entity in 2025. It does not verify every technical or community claim in the signed note.

QIX's relevance to Morin's profile is not nostalgia. It represents the local-exchange model from which his later operating argument begins. A local IXP can bring networks into the same switching environment, reduce dependence on distant interconnection points, and create a forum where operators address regional needs.

Locality alone is not a guarantee of resilience or legitimacy. Traffic can remain geographically close while depending on shared facilities, power, transport, or software. A locally governed exchange can still have unclear records, weak change control, or limited public evidence support. A national platform can improve redundancy while centralizing authority. Geography is one property of the system, not a complete evaluation.

Morin's account identifies another limit: volunteer effort and fragmented local organizations may struggle as expectations rise. Members may want extended support, higher-capacity interfaces, more locations, private connectivity, cloud access, or consistent procedures. Regulatory and customer requirements can add documentation and response obligations that a small board cannot absorb informally.

The premise is plausible, but it remains an attributed diagnosis. Public evidence in the accepted package does not audit QIX's staffing levels, incident history, capital requirements, or member satisfaction. It does not prove that the older model failed. It records the reasoning Morin uses to support a transition.

That difference matters. Infrastructure transitions are often narrated as inevitable. The old system is called too small, the new one scalable, and the change is presented as progress. A reality-layer account asks what changed in the responsibility map.

Who operates the switches after consolidation? Which policies become common? Which decisions remain local? How are member votes translated into technical priorities? Which services share staff or failure domains? What happens if one region needs a different operating rule? Can records show which entity owns each obligation?

The signed CANIX note says local committees remain part of the structure. That is relevant as an organizational intention. It does not by itself prove that local authority is effective. The evidence needed would include bylaws, decision records, service documentation, member processes, and observable operations over time.

Morin's QIX role supplies historical continuity. It also creates accountability for explaining what is preserved and what is changed. The value of continuity is not the retention of a familiar name. It is the preservation of operational knowledge and member control while the platform evolves.

Consolidation as an Operating Decision

CANIX's signed operating note describes QIX and other Canadian exchanges moving into a shared organization. Corporations Canada records that the current CANIX corporation resulted from an amalgamation involving the Montreal Internet Exchange and the National Capital Internet Exchange. The legal record establishes the entity transition; it does not describe the full technical integration.

Morin's note presents consolidation as a way to share infrastructure, expertise, support, and administrative capacity. Instead of multiple exchanges maintaining separate legal structures and operating systems, CANIX would provide a common platform while retaining regional participation.

This can solve genuine problems. Network operations benefit from repeatable procedures. A shared team can maintain software, monitoring, documentation, and security practices across sites. Procurement and vendor management may become less fragmented. Members connecting in more than one city may encounter a more consistent service.

Consolidation can also create new dependencies. A common management system can become a broad failure domain. A shared route-server policy can spread an error. Central support can lose local knowledge. Standard contracts may fit large networks better than small ones. A single governing entity may make it harder for a region to choose a different technical path.

The public evidence does not resolve those trade-offs. It gives the article a decision to examine. Morin and CANIX are choosing a broader operating structure because they believe separate local models cannot reliably carry future scale and obligations. That is a claim about organizational design, not just branding.

The right evaluation is not whether consolidation sounds modern. It is whether the new structure improves continuity without obscuring control. Records should identify who operates each component, who authorizes changes, how members receive notices, how failures are escalated, and how regional decisions interact with common policy.

The legal entity is one part of that record. The federal registry identifies CANIX, its active status, its directors, and its corporate history. It is authoritative for those fields. It does not certify network quality or member governance. A corporation can exist cleanly on paper while its technical system remains difficult to operate; a strong technical community can also suffer when legal and administrative records fall behind.

Morin's role as chair and director places him on the governance side of that boundary. His earlier account of hands-on BGP and IPv6 work places him close to the operational side. The profile becomes interesting where those layers meet.

An exchange board does not configure every session. Engineers do not decide every legal obligation. A durable IXP needs a way for operational evidence to influence governance and for governance decisions to remain implementable by the people running the platform.

The consolidation decision can be judged over time through that interface. Does the organization publish clear service definitions? Are technical changes tied to accountable owners? Do members retain meaningful choices? Does a broader platform make support more reliable without turning local networks into passive customers of another centralized service?

Those questions are not accusations. They are the normal accountability questions created by scale.

The 24/7 NOC Claim

Morin's signed CANIX note says the new operating model includes a network operations centre available around the clock. This is one of the clearest examples of a claim that should be both attributed and unpacked.

A 24/7 NOC can mean several things. It may mean staff are physically present at all times. It may mean an on-call rotation receives alerts and customer requests. It may involve an external provider handling first-line response. It may cover only particular services or severity levels. The accepted public sources do not define the implementation.

The operational value also depends on more than availability. A responder needs current documentation, monitoring, access, escalation authority, and a way to distinguish an exchange fault from a member, facility, transport, or routing-policy issue. A telephone number answered at night is not the same as a functioning incident process.

For smaller networks, responsive support can materially change the usability of an exchange. A small team may encounter a failed session or reachability anomaly without having another engineer available for peer review. An exchange operator who can verify the shared fabric, route server, port state, or maintenance status can narrow the problem quickly.

The support relationship must not remove the member's control. The exchange should not silently change a member's routing policy or treat support access as permission to operate the member network. Clear boundaries protect both sides.

Morin's ARIN remarks and CANIX note align on this point. The earlier statement described reducing the practical burden on smaller networks. The later note proposes a shared support structure. That is a coherent decision chain, even though the public evidence does not measure the result.

The record also raises continuity questions inside CANIX. A NOC depends on staffing, training, tooling, and handoffs. If expertise remains concentrated in one or two people, a round-the-clock label may mask a fragile escalation path. If procedures are too rigid, responders may be unable to act when an unusual regional condition occurs.

Observable evidence could eventually include service-level documentation, maintenance notices, status records, incident reviews, or member reports. None is required to establish the current article's bounded claim. The article can say that Morin has publicly advanced a 24/7 NOC as part of the operating model. It cannot say the NOC has achieved a particular response time or reliability level.

This is the difference between announcing continuity and demonstrating it. The announcement establishes an intended control. Running operations produce the evidence.

400G, Layer 3, and Multiservice Ports

The same signed note says new CANIX sites support 400G, Layer 3 functions, and multiservice ports. These phrases can sound like a feature list, but each changes the exchange's operating surface.

A 400G-capable interface can accommodate networks with high capacity requirements and can reduce the number of lower-speed links needed for some designs. Capability does not equal utilization. It does not reveal how much capacity is deployed end to end, whether the fabric is oversubscribed, or what redundancy exists. The article should not convert a port speed into a performance result.

Layer 3 services can help networks that need more than a shared Layer 2 peering fabric. They may provide IP services or other routing functions. They also place more policy and operational responsibility on the exchange operator. The organization is no longer only switching frames between members; it may participate more directly in routing decisions or service delivery.

Multiservice ports can simplify physical connectivity by carrying different logical services over one interface. That can reduce cross-connect costs and deployment friction. It can also make fault isolation and contract boundaries more complex. A port failure may affect several services. Configuration errors can cross logical boundaries if controls are weak.

Morin's note frames these capabilities as tools to make peering and interconnection easier. That objective connects to his earlier small-network focus. A network that cannot coordinate many vendors or circuits may benefit from a single platform. The platform then assumes responsibility for making the layers legible.

Legibility requires accurate service records. A member should be able to identify which VLAN or virtual circuit carries which relationship, who controls the route policy, what redundancy applies, and which party owns an incident at each layer. The exchange's inventory and configuration systems become part of the accountability mechanism.

This is where a running-code principle becomes practical. Service names and diagrams are not enough. The configured platform must correspond to the records. If the portal says one VLAN is active while the network implements another, the abstraction fails. If a port is marketed as redundant while both paths share a hidden dependency, the service boundary is misleading.

The accepted evidence does not audit CANIX's records or architecture. It establishes that Morin publicly associates the broader platform with these concrete capabilities. That is sufficient to explain the operating stakes without rating the implementation.

The capabilities also complicate governance. Members may disagree about whether an exchange should offer IP transit-like functions, cloud access, private services, or colocation. Some may see these services as necessary for sustainability; others may worry that they blur the exchange's neutral role.

There is no universal answer in the source package. The accountability requirement is clearer: the organization should distinguish public peering from other services, disclose the operating relationship, and avoid using community language to hide commercial or technical dependencies.

Compliance Without Permission Theater

Morin's signed note argues that Internet exchanges will face growing compliance demands and that small volunteer-run organizations may struggle to respond. The Global Peering Forum agenda independently confirms that his current public presentation topic includes critical compliance and service needs.

The evidence supports the existence of that operating concern. It does not establish which legal requirements apply to every CANIX service, predict a specific regulatory action, or prove that consolidation guarantees compliance.

Compliance can improve infrastructure when it forces accurate records, defined responsibility, security controls, and continuity planning. It can also become permission theater: a growing layer of forms and approvals that looks rigorous while leaving the running system poorly understood.

An IXP sits in a sensitive position because it enables networks to exchange traffic but does not necessarily control the content or the member networks. Rules written for carriers, data centres, cloud providers, or critical systems may not map cleanly to an exchange platform. The organization needs legal interpretation and technical clarity before turning a broad obligation into a configuration or member requirement.

Morin's proposed shared structure may make that work easier. A national organization can maintain common policies, assign owners, and gather expertise. It may be better positioned than several small boards to respond to audits, security questionnaires, or government requests.

Central capacity is not automatically legitimate authority. The exchange should not expand its control over members simply because it can describe the expansion as compliance. Every new data field, access requirement, monitoring function, or enforcement action needs a defined purpose and boundary.

The public article cannot evaluate CANIX's internal compliance program. It can identify a useful standard. Compliance should preserve operational continuity and accurate accountability without converting the exchange into an unnecessary gatekeeper.

That standard follows from the technical role of an IXP. The exchange coordinates a shared environment. It needs rules sufficient to keep that environment secure and reliable. It should not claim sovereignty over the autonomous networks connected to it.

Morin's own earlier language about making processes accessible to smaller networks creates an additional test. If compliance requirements become too complex, smaller members may be pushed back toward transit-only arrangements or intermediaries. A shared platform should reduce that burden where possible, not merely centralize it.

The outcome will depend on implementation. Clear templates, documented responsibilities, proportionate controls, and transparent change processes can help. Vague authority and opaque enforcement can harm the same networks the organization says it wants to support.

The article therefore treats compliance as an operating constraint, not a badge of maturity.

Community Language Is Not a Technical Result

CANIX describes itself as a nonprofit serving the Internet community. Morin's signed note emphasizes collaboration, local committees, and the continuation of regional exchange traditions. These are relevant governance claims. They should not be treated as evidence that the platform is automatically neutral, representative, or technically sound.

Community ownership can create useful incentives. Members may elect directors, participate in committees, and share knowledge. A nonprofit structure can reduce pressure to maximize short-term returns. Local operators can identify practical needs that a distant provider might miss.

The legal form cannot guarantee those outcomes. Participation may be uneven. Larger networks may have more time and influence. Board processes may be difficult for small members to follow. Technical decisions may remain concentrated among a few specialists. An organization can call itself community-led while ordinary members have little visibility into changes.

The same caution applies to geographic language. Keeping traffic within a country or region can reduce path length in some cases and avoid unnecessary detours. It does not automatically improve security, privacy, resilience, or cost. The result depends on routes, facilities, transport, contracts, and the actual location of services.

Morin's operating record is most persuasive when it stays close to concrete problems: a small network needs help with BGP, an exchange needs staffed support, a site needs sufficient port capability, or a shared platform needs clear service boundaries. It is less persuasive when broad language about national strength or community interest substitutes for evidence.

This does not mean the values are irrelevant. It means they must be expressed through operating rules. If local participation matters, decision records should show how local input changes outcomes. If neutrality matters, service and conflict policies should make it observable. If access for smaller networks matters, pricing, documentation, onboarding, and support should reflect it.

Corporations Canada's record contributes a useful reality check. It identifies the legal entity and directors. That record is not a certificate of technical legitimacy. It is one layer of accountability. CANIX's public service descriptions are another. The running exchange, member relationships, and support performance form additional layers.

Morin sits across those layers as a named chair, director, operator, and public speaker. The article can examine that position without accepting the organization's preferred story.

A mature exchange should be able to tolerate that distinction. Community language can describe an aspiration. Running operations and records determine whether the aspiration has practical meaning.

Governance as Recordkeeping

Internet exchange governance is sometimes discussed as representation and sometimes as engineering. In practice, it also depends on disciplined recordkeeping.

The federal corporate registry records the legal entity, directors, and amalgamation history. CANIX's site records board roles, locations, and service categories. Technical systems should record ports, VLANs, route-server participation, maintenance, incidents, and authorization. Member agreements record rights and obligations.

These records serve different purposes. None should be used to prove what belongs to another layer. A director listing does not prove technical control. A port inventory does not prove legal authority. A signed operating note does not prove measured performance. A BGP session does not prove member consent beyond its configured relationship.

The value comes from alignment. The entity operating the service should match the contracts and public descriptions. The people authorized to approve changes should correspond to governance records. The technical inventory should correspond to the running configuration. Incident ownership should correspond to the service boundary.

Consolidation increases the number of relationships that need to remain aligned. A shared platform can simplify the top-level structure while multiplying internal dependencies. Regions, sites, members, services, and vendors need stable identifiers and change histories.

Morin's public plan can be read as an attempt to professionalize that system. A NOC, common platform, and standardized services can make records more consistent. They can also hide local detail if the abstraction is too broad.

The operating challenge is to standardize what should be common without erasing what needs to remain specific. A maintenance procedure may be shared while facility contacts differ. A route-server policy may be common while peering LANs have separate prefixes and failure domains. A contract may use a common format while regional access arrangements vary.

Recordkeeping also supports portability. A member should be able to understand its relationships well enough to change capacity, add a site, stop a service, or move to another arrangement. An exchange that makes exit or change opaque creates dependence even if it is nonprofit.

This is a practical form of accountability. It does not rely on moral language. It asks whether the records allow operators to coordinate, verify authority, recover from failure, and change relationships without losing control.

Morin's earlier ARIN focus on making Internet processes usable for small networks fits this standard. Small teams benefit when the shared system records complexity accurately instead of pushing that complexity into private knowledge.

The article cannot show whether CANIX has achieved that alignment. It can show why the alignment is the right measure.

What Scale Should Preserve

Scale is not only the ability to carry more traffic or connect more sites. It is the ability to add relationships without losing operational understanding.

For CANIX, scale may involve additional cities, members, port speeds, services, facilities, and support obligations. Each addition creates identifiers, dependencies, and possible failure paths. A national label does not make those relationships simpler by itself.

The earlier QIX model offered a relatively clear local context. Networks connected in Montreal to exchange traffic. As the platform expands, a member may connect in one city, use transport to another, consume several logical services, and rely on shared support or routing infrastructure.

The organization should preserve the member's ability to answer basic questions. Where does traffic enter the platform? Which component makes each routing decision? Which services share a failure domain? Who can authorize a change? What happens when a region is isolated? Which records are authoritative?

Morin's signed note presents local committees as a way to preserve regional participation. That may help if the committees have defined authority, current technical information, and a path into common governance. A committee that exists only as a label cannot preserve local control.

The platform also needs to preserve the autonomy of connected networks. An IXP works because autonomous systems choose relationships. A common exchange can make those relationships easier, but it should not collapse them into a single hidden policy.

Route servers illustrate the balance. They can reduce the number of bilateral sessions and simplify onboarding. Members still need clear control over import and export policy. The route server should not make the exchange the sovereign of member routing.

Private VLANs and multiservice ports create a similar balance. Shared infrastructure can reduce cost. Records and isolation must remain accurate enough that one service does not become an unexplained dependency of another.

Support can preserve continuity when it transfers knowledge and provides reliable escalation. It can undermine autonomy if members cannot diagnose their own relationship or if the exchange becomes the only party capable of interpreting it.

Morin's public record connects these questions to a person because he has described both the small-network burden and the larger platform response. The credibility of the response depends on whether scale preserves agency.

That is a more demanding standard than growth. It requires the system to remain understandable as it becomes broader.

What the Public Record Leaves Unknown

The accepted evidence leaves important questions unanswered. It does not provide CANIX's network topology, route-server configuration, redundancy design, service-level targets, incident history, or staffing model.

The signed note says CANIX has or plans a 24/7 NOC, 400G-capable sites, Layer 3 functions, and multiservice ports. It does not define coverage, deployment status at every location, capacity engineering, or observed performance.

The public CANIX site lists services and locations. It does not show which service is available at each site under which technical and contractual conditions. It does not expose every vendor or facility dependency.

The federal registry establishes corporate facts. It does not show how board decisions are made, how regional committees operate, how conflicts are managed, or how members influence technical priorities. This article deliberately excludes postal addresses and other contact details that are irrelevant to the analysis.

The GPF agenda shows that Morin was assigned a topic about scale, compliance, and service needs. It does not provide a transcript or independent validation of the presentation's claims.

The ARIN transcript records Morin's 2021 account of helping small networks with IPv6, BGP, and peering. It does not name deployments or provide technical results. It also reflects his role and priorities at that time, not a complete current biography.

The records do not establish that Morin alone designed CANIX, that every participating exchange agreed with the transition, or that the model is superior to every local alternative. They do not establish uptime, security quality, traffic growth, latency improvement, or member savings.

These limits are not a weakness in the article. They define its proper scope. The evidence is sufficient to connect Morin to a real operating problem and a concrete organizational response. It is not sufficient for an audit or performance rating.

Future evidence could include technical service documentation, public route-server policy, membership governance records, status history, or independent network observations. Such evidence would need to be interpreted at its own layer.

A route observation could show paths, not internal governance. An incident report could show a failure and response, not general reliability. A board record could show a decision, not its technical outcome.

The current public record supports a source-bounded profile. It does not support certainty beyond it.

The Reality Layer of Exchange Operations

Internet exchanges are easy to describe as neutral places where networks meet. The running system is more complicated. Ports have states. Optics fail. VLANs are provisioned. Route-server policies change. Facilities schedule maintenance. Members announce unexpected routes. Support teams need evidence to separate a shared-fabric problem from a member configuration problem.

Morin's public record is relevant because it stays close to that operating layer. In the ARIN forum, he talked about small networks navigating BGP and IPv6. In the CANIX note, he described support, capacity, Layer 3 functions, and a common platform. At GPF, his listed topic joined scale, compliance, and service needs.

Those records form a coherent public trajectory. They do not prove success. They show the problems Morin chooses to address and the operating model he advocates.

The reality-layer test is concrete. A registry or governance record should identify accountable relationships. A service description should correspond to running capability. A NOC claim should correspond to an actual response process. A community claim should correspond to observable member authority. A consolidation claim should correspond to continuity across the transition.

The exchange is not sovereign over connected networks. It is a shared coordination system. Its legitimacy comes from accurate service boundaries, operational competence, member choice, and the ability to maintain continuity, not from geography or nonprofit status alone.

Morin's small-network emphasis gives the test a practical beneficiary. Smaller operators should not need a large policy team simply to understand the exchange. They should still retain responsibility for their routes and systems. The exchange succeeds when it removes avoidable coordination friction without hiding essential control.

CANIX's broader platform may make that possible. Shared staff and systems can provide more consistent support. Higher-capacity and multiservice options can reduce deployment barriers. Common governance can align legal and technical responsibilities.

The same platform may create central dependencies. The only responsible conclusion at this stage is that the design should be judged by its records and running behavior.

Morin's significance is therefore not a title. It is the continuity between a public account of hands-on routing access and a later attempt to redesign the institution around that access problem.

The next evidence should come from operation, not promotion.

Sources