Robert Hansel & Benjamin Ihrig – InnerSource Contribution Insights #InnerSourceSummit2025

Robert Hansel & Benjamin Ihrig – InnerSource Contribution Insights #InnerSourceSummit2025 🔗 **Source:** [Robert Hansel & Benjamin Ihrig – InnerSource Contribution Insights #InnerSourceSummit2025](https://www.youtube.com/watch?v=GniF0FkJM9A) > [!note]+ Document Notes > tags: [innersource, enterprise-collaboration, data-analytics] > author: InnerSource Commons > source: YouTube > reference: https://www.youtube.com/watch?v=GniF0FkJM9A > date created: 2024-12-19 09:30 > > # 🏷️ Title: Robert Hansel & Benjamin Ihrig – InnerSource Contribution Insights #InnerSourceSummit2025 >  ## 🧰 Action Inbox - [H1] Implement organizational distance metrics for inner source identification — 08:45 #measurement - [H2] Create heat maps for cross-entity collaboration tracking — 11:23 #analytics - [H3] Set up automated data collection from Git, directories, and ERP systems — 04:12 #infrastructure ## 🧭 Table of Contents - [📝 Summary](#-summary) - [📖 Main Content](#-main-content) - [🗒️ Notes](#️-notes) - [📚 References](#-references) - [🗣️ Paraphrase](#️-paraphrase) - [📰 Newsletter](#-newsletter) - [✅ Action Items](#-action-items) - [🧪 Hypotheses](#-hypotheses) - [📊 Methodology](#-methodology) - [❓ Questions](#-questions) - [🔮 Future Research](#-future-research) - [💡 Implications](#-implications) - [🧠 Learning Styles Assessment](#-learning-styles-assessment) - [📊 Accessibility Assessment](#-accessibility-assessment) - [🌍 DEIB Assessment](#-deib-assessment) - [✅ Logical Validity](#-logical-validity) - [🛡️ DARVO Analysis](#-darvo-analysis) - [🎯 Conclusion](#-conclusion) ### 🔗 References - [InnerSource Commons](https://innersourcecommons.org/) - [SAP Open Source](https://opensource.sap.com/) ## 📝 Summary - SAP and Bosch shared their approaches to measuring inner source collaboration effectiveness through data analytics - Both companies use different methods to identify when cross-organizational collaboration breaks down silos - SAP focuses on bottom-up measurement of inner source activity using organizational distance metrics - Bosch emphasizes regulated inner source with heat maps tracking multi-entity contributions - Key insight: Companies need tailored metrics based on their organizational structure and goals ## 📖 Main Content Two enterprise open source program offices presented their data-driven approaches to measuring inner source success. Benjamin Ihrig from SAP and Robert Hansel from Bosch explained how they collect and analyze commit data, organizational information, and collaboration patterns to answer stakeholder questions about inner source effectiveness. SAP takes a bottom-up approach, seeking to identify genuine inner source commits by measuring organizational distance between contributors and project teams. They calculate the hierarchical distance in their org structure between a contributor and the most active project committer, with larger distances indicating more likely inner source activity. Bosch, with over 400 legal entities and 400,000 employees, uses a more regulated approach requiring specific licenses and permissions for inner source participation. They create heat maps showing contributions across organizational entities, considering projects with five or more contributing legal entities as successful inner source initiatives. Both companies emphasized that while companies may appear similar externally, internal structures require customized measurement approaches. The core question remains consistent: Does collaboration happen outside of mandated top-down partnerships, genuinely breaking down organizational silos? ## 🗒️ Notes #### ⏱️ [00:00](https://youtu.be/GniF0FkJM9A?t=0) — [H1] Speaker introductions and company contexts - Key idea: Benjamin from SAP OSPO, Robert from Bosch OSPO presenting joint insights - Evidence: "We've been brought together by GM and Georg because we both applied with a rather similar topic" #### ⏱️ [01:15](https://youtu.be/GniF0FkJM9A?t=75) — [H2] Bosch organizational complexity - Key idea: Bosch has over 400,000 employees across 400+ legal entities - Evidence: "We are roughly a little bit more than 400,000 employees worldwide and we have more than 400 different legal entities" #### ⏱️ [02:30](https://youtu.be/GniF0FkJM9A?t=150) — [H3] Company similarities and differences analogy - Key idea: Companies appear similar externally but have different internal architectures - Evidence: "If you look at companies from the outside, companies have many similarities... but if you dig deeper, you'll see that there's still similarities like CPUs, GPUs, but some have Intel, some have others" #### ⏱️ [03:45](https://youtu.be/GniF0FkJM9A?t=225) — [H2] SAP's bottom-up inner source approach - Key idea: SAP wants to measure actual inner source activity happening organically - Evidence: "Inner source in SAP is a very bottom-up approach and we want to see how much it pays off, how much inner source activity is really happening" #### ⏱️ [04:12](https://youtu.be/GniF0FkJM9A?t=252) — [H3] Bosch's regulated inner source model - Key idea: Bosch has more regulation and restricted participation in inner source - Evidence: "We have way more regulation around it than SAP and we have a restriction who can actually join inner source inside of Bosch" #### ⏱️ [05:00](https://youtu.be/GniF0FkJM9A?t=300) — [H1] Stakeholder personas and their questions - Key idea: Different roles ask different questions about inner source effectiveness - Evidence: "There are different people - Martin the manager, Amelia the engineer, the developer, the inner source expert" #### ⏱️ [06:15](https://youtu.be/GniF0FkJM9A?t=375) — [H2] Manager questions about inner source - Key idea: Managers want to understand costs, benefits, and success measurement - Evidence: "Manager typically asks what is inner source, what does it cost, how do we measure success, what are the actual benefits" #### ⏱️ [07:30](https://youtu.be/GniF0FkJM9A?t=450) — [H3] Engineer and developer information needs - Key idea: Engineers seek project information, contacts, and networking opportunities - Evidence: "Engineer is interested in where could I find information, who's involved in a project, who could I contact" #### ⏱️ [08:45](https://youtu.be/GniF0FkJM9A?t=525) — [H1] Core question about breaking down silos - Key idea: The fundamental question is whether collaboration happens outside mandated partnerships - Evidence: "Key question always for us is does collaboration happen outside of the regular collaboration which is forced from top level" #### ⏱️ [09:30](https://youtu.be/GniF0FkJM9A?t=570) — [H2] Data sources: Git commits and metadata - Key idea: Analysis starts with commit data including author, timestamp, message, and repository info - Evidence: "The lowest entity is the commit... has metadata like who is the author, when did it happen, the message" #### ⏱️ [10:45](https://youtu.be/GniF0FkJM9A?t=645) — [H3] Repository-level data collection - Key idea: Repository data includes maintainers, contributors, licenses, languages, and dependencies - Evidence: "Repository contains information about maintainers, contributors, dedicated license internally, different languages being used" #### ⏱️ [11:15](https://youtu.be/GniF0FkJM9A?t=675) — [H2] External data integration - Key idea: Companies integrate directory services and ERP systems for comprehensive analysis - Evidence: "Add other data sources like Microsoft Entra ID... employee information like location, department, cost center" #### ⏱️ [12:30](https://youtu.be/GniF0FkJM9A?t=750) — [H1] Organizational insights from combined data - Key idea: Combined data enables analysis of programming languages, technologies, and developer engagement by business unit - Evidence: "You can start answering questions like in my business unit what are the top programming languages, what technologies are being used" #### ⏱️ [14:00](https://youtu.be/GniF0FkJM9A?t=840) — [H2] SAP's privilege-based inner source identification - Key idea: SAP identifies inner source commits by analyzing repository privileges and team membership - Evidence: "If someone has write access or more we perceive him to be part of the maintaining team, everyone with less privileges is likely to be an inner source committer" #### ⏱️ [15:30](https://youtu.be/GniF0FkJM9A?t=930) — [H3] Organizational distance methodology - Key idea: SAP calculates hierarchical distance between contributors and core team members - Evidence: "We take the most active committer, say he's part of core team, then calculate distance in org structure - larger distance means more likely inner source commit" #### ⏱️ [16:45](https://youtu.be/GniF0FkJM9A?t=1005) — [H2] Bosch's heat map approach - Key idea: Bosch creates heat maps showing contributions across legal entities to identify silo breakdown - Evidence: "We build a heat map per repository looking at organizational entities - the more entities contributing, the more likely we see breakdown of silos" #### ⏱️ [17:30](https://youtu.be/GniF0FkJM9A?t=1050) — [H3] Success threshold for cross-entity collaboration - Key idea: Bosch considers five or more legal entities contributing as successful inner source - Evidence: "Two or three legal entities is quite common because of contracting, but five or more is rare - at that threshold you can assume it's real inner source" ## 📚 References - [InnerSource Commons Summit 2025](https://innersourcecommons.org/) - [SAP Open Source Program Office](https://opensource.sap.com/) - [Bosch Digital](https://www.bosch-digital.com/) ## 🗣️ Paraphrase - Two major enterprises shared complementary approaches to measuring inner source effectiveness through data analytics, with SAP focusing on organic bottom-up collaboration measurement and Bosch emphasizing regulated cross-entity engagement tracking - Both companies collect similar foundational data from Git repositories and organizational systems but apply different analytical methods based on their unique corporate structures and goals - The core challenge both organizations address is determining when genuine collaboration occurs outside of mandated partnerships, indicating successful breakdown of organizational silos ## 📰 Newsletter **Enterprise Inner Source Analytics: Lessons from SAP and Bosch** Two major enterprises recently shared their data-driven approaches to measuring inner source success at the InnerSource Commons Summit. While both companies collect similar foundational data from Git repositories and organizational directories, they've developed distinct analytical methods tailored to their unique structures. SAP takes a bottom-up approach, using organizational distance metrics to identify genuine cross-team collaboration. Meanwhile, Bosch with its 400+ legal entities focuses on heat maps tracking multi-entity contributions, considering projects with five or more contributing entities as successful inner source initiatives. The key insight for community builders: while measurement frameworks may look similar externally, internal implementation requires deep customization based on your organization's specific structure and collaboration patterns. **Key Takeaways:** - Combine Git data with organizational directories for comprehensive analysis - Develop metrics that align with your company's unique structure - Focus on measuring genuine collaboration beyond mandated partnerships Resources: [InnerSource Commons](https://innersourcecommons.org/), [InnerSource Patterns](https://patterns.innersourcecommons.org/), [OSPO Alliance](https://ospo.zone/) ## ✅ Action Items #### OSPO - Policy & Governance: Develop inner source measurement frameworks aligned with organizational structure - Software Inventory: Integrate Git repositories with organizational directories and ERP systems - Developer Support: Create dashboards showing cross-organizational collaboration patterns - Community Engagement: Use collaboration data to identify successful inner source champions - Strategic Alignment: Establish KPIs for silo breakdown and cross-entity engagement #### Community - Maintainers: Track contributor diversity across organizational boundaries - Contributors: Provide visibility into cross-team collaboration opportunities - Sponsors: Demonstrate inner source ROI through collaboration metrics #### DevRel - Developer advocacy: Use inner source success stories from collaboration data - Education: Create workshops on measuring and improving cross-organizational collaboration - Developer experience: Build tools that surface cross-team collaboration opportunities - Developer marketing: Highlight successful inner source projects breaking down silos ## 🧪 Hypotheses - Organizations with greater hierarchical distance between contributors and maintainers demonstrate more effective inner source practices - Projects with contributions from five or more organizational entities indicate successful silo breakdown - Automated privilege-based detection can accurately identify inner source contributions - Heat map visualization of cross-entity contributions correlates with organizational culture change ### 📊 Methodology - Implement organizational distance calculations using corporate directory APIs - Create repository heat maps tracking contributor entity diversity over time - Analyze privilege patterns in GitHub repositories to classify contribution types - Survey contributors to validate automated inner source identification accuracy ### ❓ Questions - How do different organizational structures affect optimal inner source measurement approaches? - What threshold of cross-entity contributions indicates genuine cultural transformation? - Can automated detection replace manual inner source project classification? ### 🔮 Future Research - Comparative analysis of inner source measurement effectiveness across industries - Machine learning approaches to inner source contribution classification - Correlation studies between inner source metrics and business outcomes ### 💡 Implications - Practice: Companies need customized inner source measurement approaches based on organizational structure - Policy: Inner source governance should account for varying collaboration patterns across business units - Security: Cross-organizational code sharing requires careful privilege management and access controls - Education: Training programs should emphasize organizational context in inner source implementation - Funding: Investment in inner source tooling should prioritize analytics and measurement capabilities ### 🧠 Learning Styles Assessment Reading level estimate: Flesch-Kincaid Grade Level 12 #### 👀 Visual learners - Present: Organizational charts, heat maps, repository contribution diagrams - Improve: Interactive dashboards showing collaboration patterns, network visualizations #### 🔊 Aural learners - Present: Conference presentation with speaker explanations - Improve: Podcast discussion on measurement approaches, recorded stakeholder interviews #### 🗣️ Verbal learners - Present: Clear definitions of inner source terms, stakeholder persona descriptions - Improve: Glossary of measurement terminology, narrative case studies #### 🧩 Kinesthetic learners - Present: Data collection and analysis workflows - Improve: Hands-on workshop building collaboration dashboards, interactive data exploration #### 🧠 Logical learners - Present: Step-by-step measurement methodologies, clear decision trees - Improve: Algorithmic flowcharts for contribution classification, statistical validation frameworks #### 🧑‍🤝‍🧑 Social learners - Present: Joint presentation format, stakeholder perspective discussions - Improve: Community of practice for inner source measurement, peer learning sessions #### 🧘‍♀️ Solitary learners - Present: Detailed technical documentation, self-contained examples - Improve: Self-assessment tools for inner source maturity, independent analysis exercises #### 🌿 Naturalistic learners - Present: Ecosystem thinking about organizational collaboration - Improve: Biological analogies for organizational health, sustainability metrics ### 📊 Accessibility Assessment #### Screen reader semantics - Present: Clear heading hierarchy, descriptive slide content - Improve: Alt text for organizational charts and heat maps #### Keyboard navigation - Present: Standard conference presentation format - Improve: Accessible dashboard interfaces for data exploration #### Contrast and readability - Present: High contrast presentation slides - Improve: Ensure data visualizations meet accessibility standards #### Text resizing and reflow - Present: Scalable presentation format - Improve: Responsive dashboard layouts supporting zoom #### Media and captions - Present: Video presentation available - Improve: Professional transcript, slide synchronization with audio #### Forms and interactions - Present: Q&A session format - Improve: Accessible survey tools for stakeholder feedback ### 🌍 DEIB Assessment #### Cultural breadth - Present: Global enterprise perspectives from European companies - Improve: Include diverse geographic and company size examples #### Inclusive language - Present: Professional, jargon-explained terminology - Improve: Plain language summaries, cultural context considerations #### Neurodiversity support - Present: Structured presentation with clear sections - Improve: Multiple content formats, processing time accommodation #### Representation - Present: Two different company approaches, diverse stakeholder personas - Improve: Include perspectives from various organizational roles and levels --- ## Full Transcript [00:06] Yeah, nice to see you here in Berlin. [00:07] Um, yeah, quick words to myself. I'm [00:11] Benjamin from the SAP open source [00:13] program office. Um, you want to [00:16] introduce yourself? [00:17] >> Yeah, sure. I'm Robert. I'm with the [00:19] OPOS of Bosch. [00:22] >> Yeah. So, uh, we've been brought [00:24] together. Oh, no. We have to talk about [00:26] Bosch quickly. [00:26] >> Yeah, we slide first. So just like we [00:30] are roughly a little bit more than [00:32] 400,000 employees worldwide. Um and we [00:35] have more than 400 different legal [00:38] entities. So like from an organizational [00:39] point of view quite a complex structure. [00:41] We are active in in certain areas like [00:43] like household appliances, cars and and [00:47] other things. You can see it all here. [00:49] And like the the OSOS is currently [00:51] located in the Bosch digital which [00:53] >> morning in progress. [00:55] >> Okay. Nice. uh in in the Bosch Bosch [00:58] digital branch which is kind of like the [01:00] central IT infrastructure service [01:02] provider for all of the different legal [01:04] entities of Bosch. [01:07] >> Yeah. And Peter already talked a lot [01:09] about SAP, so I will skip on that one. [01:12] If you have questions, you can come out [01:13] to anyone with the the screenshots with [01:15] SAP on it. Um yeah, we've been brought [01:18] together by GM and Georg because we both [01:22] oblied with a rather similar topic to [01:25] the summit and they thought well why not [01:27] do a talk together. Um and then uh [01:30] Robert and I sat together and found that [01:32] many companies have many similarities [01:34] like this computer here. It's not a dual [01:37] toaster, it's a computer. Um so yeah, if [01:40] you look at companies from the outside, [01:42] companies have many many similarities, [01:45] right? especially when you talk about [01:46] stuff like inner source. Um, but if you [01:48] dig deeper, um, you'll get to the [01:51] interns of of the computer and the [01:53] company and you'll see that there's [01:56] still similarities like we still have [01:58] CPUs, we have GPUs and all that like a [02:00] main board, but in detail you'll find [02:03] that some companies have Intel, some [02:06] have others. Um and as such also inner [02:10] source differs and for example we'll be [02:13] talking about data about insights it's [02:15] also different on what we want to gather [02:17] out of the the of the data we acquire um [02:21] for example inner source in SAP is a [02:23] very bottomup approach and we want to [02:25] see how much it pays off what we're [02:26] doing how much inner source activity um [02:29] is really happening across our company [02:32] um so what we are up to is look at the [02:35] activity in our GitHubs and actually see [02:38] what of these activities are inosour [02:41] activities. [02:44] Yeah. And within Bosch we have kind of [02:47] like inner source as well. But I would [02:49] say like from the talks to Benjamin we [02:51] have way more regulation around it I [02:53] would say than than SAP. And so we have [02:55] certain other things which make [02:57] innosource for us innov. [03:03] And we also have a restriction who can [03:05] actually join inner source inside of [03:07] Bosch because not every employee inside [03:09] of the company can actually join. So [03:10] there are slight um differences between [03:13] how we approach that. [03:16] Yeah. And so we thought okay um when [03:18] we're talking about data why are we [03:20] talking about data and why do you [03:22] collect data? It's usually because you [03:23] want to answer questions. And so there [03:25] are a couple of different people and [03:27] maybe you can relate to one of them. [03:29] It's kind of like like for example we [03:30] have Martin the manager, we have Amelia [03:32] the engineer and then we have the [03:34] developer and and the innosource expert [03:36] for example like the both of us and then [03:38] maybe also an analyst and they are all [03:40] interested in in different things about [03:41] innosource and then it's kind of like a [03:42] manager typically first question what is [03:44] insource if he doesn't know yet. Um then [03:47] the next question is obviously what what [03:49] does it cost? How do we measure cost? [03:50] How do we measure success and and what [03:52] are the actual benefits and these kind [03:54] of things, right? [03:57] >> Yeah. And for example, the engineer uh [03:59] is is rather involved uh in where could [04:02] I find information on something? Who's [04:04] involved in a project? Who could I ask? [04:06] Who could I contact? Um is a project or [04:11] some some topic relevant for my domain? [04:13] Can I expand my network within the [04:15] company? Um relatively similar questions [04:19] does have the developer, right? See, so [04:21] he's also interested in where can you [04:23] find something. Um, but he's maybe also [04:26] interested in which technology is being [04:28] used in in a project and how others do [04:31] things that he might be necessary to do [04:34] as well. Um, and maybe also the [04:36] questions if my organization is already [04:39] engaged uh in whatever topic. [04:42] Um, the the inner source expert who many [04:45] of us are uh is rather interested does [04:48] whatever we do within the company pay [04:51] off for the company. So how much is [04:53] really happening? Uh is insos growing? [04:56] Who is involved in which uh topics or [04:58] projects? Um are we actually breaking [05:00] down the silos right that big macha many [05:03] companies have? Um and yeah do my [05:06] measures really work uh to to grow in a [05:10] sourcing companies. [05:13] Yeah. And then maybe the analyst is also [05:15] interested in kind of like on a broader [05:17] scope maybe on a legal entity level or [05:19] even on the global uh company level is [05:22] kind of like what are key technologies [05:23] the company's working on what is kind of [05:25] like maybe what's the global footprint. [05:27] So where can we see engagement in [05:28] innosource around the world? Um and then [05:32] where we might want to invest not in [05:34] terms of maybe budget or money but maybe [05:36] in terms of capacity of people actually [05:38] doing stuff right and so but the key [05:41] question always for us and and for the [05:43] company is really this does [05:45] collaboration happen outside of the [05:47] regular collaboration which is kind of [05:49] like forced from top level. So managers [05:51] saying yeah you two have to work [05:52] together. Is that really? So, are we [05:54] really breaking down those silos at the [05:56] end of the day? [05:59] >> Yeah. So, let's start with a bigger [06:01] picture and we'll zoom into all of these [06:03] details. Don't worry, we'll we'll take a [06:05] closer look at that. Um, but let's start [06:07] at the very central thing uh which is [06:10] actually the the places where we keep [06:11] our code and do our work with. Um, and [06:15] the the lowest entity so to say is the [06:18] commit, right? Every might recognize [06:21] that git tree here. Um and a commit has [06:24] has several yeah metadata attached to it [06:27] like who is the author of a commit when [06:29] did it happen the message so what really [06:32] uh was part of that commit um any [06:34] additional statistics like I know maybe [06:37] amount of code or so and the repository [06:40] um it belongs to the repository is the [06:43] next greater entity um containing [06:46] information about the maintainers [06:48] contributors to a project um rather [06:51] specific to Bosch because of the the [06:53] amount of legal entities is a dedicated [06:56] license internally. An innosource [06:58] license um the the different languages [07:00] being used in a project uh its [07:02] dependencies and also the the time it [07:05] was created and things happened in the [07:08] project. [07:11] If we look away from the the [07:14] repositories the the data or the the the [07:16] code um location we can also add other [07:20] um yeah d data sources to general [07:23] analytics. Um, and that could be [07:25] directories like for example Microsoft's [07:27] Enter ID. Many of you might know that [07:29] which is the the Saka of uh Azure Active [07:32] Directory. Um, and they can actually for [07:35] example access employee information like [07:38] where they're located, uh, what [07:40] department they belong to, what cost [07:42] center they they are assigned to, um, [07:45] and also for example access the the [07:47] company's um, hierarchy and structure. [07:52] Yeah, and we at Bosch, we go also one [07:54] step further because we are also looking [07:56] into our ERP system because we have [07:58] those legal entities and we actually [07:59] need to figure out okay like which [08:01] internal organization belongs to which [08:03] legal entity and things like that. And [08:05] so we are also taking the data out of [08:07] there and what you essentially can do [08:10] with that is now you can like [08:14] pretty much sketch a broader picture of [08:16] the contributor itself. Right? So now [08:17] you have the commit data, you have the [08:19] repository where he or she was active in [08:21] and now you have also like the country [08:22] and the department but also the legal [08:24] entity and the the overall organization [08:27] and now what you can actually do is like [08:29] you can start answering all those [08:31] questions you you have seen earlier. [08:32] It's kind of like so in my business unit [08:34] what are the top like programming [08:36] languages? So so easy right? You can do [08:38] that on department level and then go up [08:39] to all the way to the whole company. You [08:41] could also see about like what are the [08:44] technologies being used? Um where might [08:46] it make sense if it's an open source [08:47] technology to maybe let people [08:49] contribute to because it's it's mission [08:51] critical for a huge chunk of your [08:53] organization. Um, you can also see kind [08:56] of like how many developers per [08:57] organizational entity are involved and [09:00] where do you might need to go into and [09:01] figure out okay if you if you have kind [09:03] of like a comparison between what are [09:06] the overall number of developers who [09:08] could engage and what are the actual [09:10] numbers who are engaged and it's kind of [09:11] like it would give you some idea where [09:13] you could do targeted um marketing for [09:16] inner source for example and then you [09:19] could like if you if you have those [09:20] historiized data you can also see that [09:22] over time and see growth over time [09:24] compared to legal entities and then you [09:26] can also all of a sudden start seeing [09:28] okay did you have this conference with [09:30] this like organizational entity a couple [09:32] weeks back or a couple months back and [09:34] can you actually see a growth path right [09:36] so is it actually really something which [09:38] is worth investing in and obviously the [09:41] global footprint so for for us it's [09:43] really interesting to see does that also [09:46] match with the the key developer [09:47] locations we have so we have a very [09:49] large global footprint inside of Bosch [09:51] but not every country is a major [09:53] development or a developing country in [09:56] terms of a lot of developers, software [09:58] engineers and and so forth are there. [10:04] >> Yeah. Uh I said before uh we SAP have [10:07] been more interested in finding out if [10:09] for example a commit is an inner source [10:11] commit. So if it was done from um let's [10:15] say foreign party to to the project [10:17] team. Um and we essentially came up with [10:19] different methods to to measure that. Um [10:22] there's stuff like we look what the [10:24] privileges of a person are in a certain [10:26] repository. So for example, if someone [10:28] has has rise access or more um we [10:31] perceive him to be part of the [10:33] maintaining team and everyone with less [10:36] um privileges is likely to be an inner [10:39] source committer. Um another option is [10:41] that you can configure a GitHub team to [10:43] contain all maintainers and everybody [10:45] who's not part of that um will then be [10:49] an inner source committer. Um both have [10:51] their their drawbacks like for example [10:53] for very big repositories mono [10:55] repositories where you have different [10:57] configurations and maybe also different [10:59] branch protections which is a GitHub [11:01] specific thing. Um there are these [11:03] things do not work too well. Yeah. Also [11:06] manually configuring a team I mean we we [11:08] all know how likely it is that we'll [11:10] take care good care of manual work. So [11:13] that team will be set up once and then [11:15] it's wrong after a certain time with [11:17] people being added or removed from the [11:19] team. Um so there's one method we try to [11:23] look into which is our or structure. [11:26] This is our or structure. Um and we well [11:29] we we have the person doing a commit [11:31] right might be max or so. And then we [11:35] try to find out to which team the the [11:38] project belongs. We do not have that [11:40] configured unfortunately in in a clear [11:43] uh concise way in SAP. So we have to use [11:46] something to find that out. Um and [11:49] currently we take the the most um active [11:52] committer to a project and say he's part [11:54] of the the core team and then we take [11:56] him in the in the O structure for [11:59] example him and him and then we [12:02] calculate the distance in between in our [12:04] o structure and then larger this [12:06] distance is then more likely it is an [12:08] inner source commit. So that's one thing [12:10] we currently try to evaluate. It's very [12:13] specific to be honest uh per projects [12:16] where we have projects that are pretty [12:18] much spread across multiple [12:19] organizations but they're still part of [12:21] a core team. Um and we have others where [12:24] it's very clear that when you're above [12:26] like four or so as a measurement um it's [12:29] more likely to be uh um in a source [12:31] commit. But yeah, as I said that's [12:33] statistics. So that's not as precise as [12:36] hard as saying yes or no. But that's [12:38] something we'll we're trying to get [12:39] along with to to measure in a source [12:42] activity in our company. [12:45] >> Yeah. And we pretty much have the same [12:48] question. It's kind of like is it really [12:51] an a successful inner source repository? [12:53] And then success in that way means kind [12:55] of is it breaking down silos? Do we have [12:57] a lot of engagement from different [12:58] entities? And so rather than doing this [13:00] organizational thing which is quite [13:02] interesting. So I learned that in the [13:04] preparation of the talk and and I I will [13:06] think about that maybe we integrate it [13:08] as well inside our company. But what we [13:10] currently do is we rather look at the [13:12] contributions and then we pretty much [13:14] build a heat map per repository and then [13:16] we look at the organizational entities [13:18] or legal entities so to speak and the [13:20] more which are contributing to a single [13:22] repository um the more likely it is that [13:25] we actually see this breakdown of silos. [13:27] If we have like two or three legal [13:29] entities, that is quite common inside of [13:31] Bosch because they have all always [13:33] collaboration going on, right? There's [13:34] contracting and things like that. But [13:36] what you rarely see is kind of like five [13:39] or more, right? And then at a certain [13:42] threshold, you can simply assume that [13:44] this is actually a real um inner source [13:47] project and really successful. [13:50] Yeah. So that is all about how we like [13:53] combine the data and what questions are [13:55] we trying to answer with that. And you [13:56] can see you can already answer a lot of [13:58] those um questions we we raised earlier. [14:02] And to some extent we can actually see [14:05] and and say yeah we see this breakdown [14:08] of silos. Maybe not to the extent you [14:10] can see on the image here. So we are not [14:11] there yet where we want to be but it's [14:13] kind of like it really starts to to be [14:15] broken down. Um [14:18] yeah and now it would be time for Q&A.