OpenStreetMap

Summary Report on OSMF Chair's Outreach Jan-early Apr 2020

Posted by apm-wa on 18 April 2020 in English (English)

Background

Shortly after the new year began, the OSM Foundation chair started contacting members of the OSM community writ large to collect information on the state of the community and project, and to assess attitudes toward the Board’s work. Most conversations were confidential in order to ensure that respondents would speak openly, frankly, and honestly (and it is the chair’s sense that virtually all of them did, and in fact some of them were quite brutally frank about the Board’s perceived shortcomings). Thus, this report will not detail “who said what”. It will, rather, tend more to aggregate viewpoints expressed during the conversations, with illustrative but unattributed quotes.

The chair began by polling members of the Board of Directors of the Foundation, and expanded to members of the Advisory Board, including both corporate members and local chapters. The latter tended to be conference calls with multiple members of those chapters. These calls generated recommendations that the chair talk to pillars of the OSM community or representative members of a tribe (e.g., software developers), so the chair reached out to those individuals as well. Another result was outreach to local communities not formally affiliated with the OSMF, and those conversations proved to be among the most fruitful. The chair held one conversation with a local community and two with corporate representatives face to face due to the happy coincidences of parties being in the same geographic location.

This effort is not over. If anything, the conversations revealed a desire for better communication between the Board and the community’s various tribes (including working groups), which can only be satiated by making the effort to reach out, to schedule calls, then just to call. Geographic coverage of the current outreach effort remains a work in progress; to date the chair has made no calls to Latin America, for example.

The Top Lines

Most conversations started with two open-ended questions: “What should the Foundation Board be focused on?” and “What do I as a Board member need to know to make OSM a continued success?” The chair sought not to direct responses in any direction, but to allow respondents to address whatever they saw as the highest priorities. That said, the chair did solicit reactions to the OSMF Board’s then-recent release of a diversity statement and formation of the Diversity and Inclusion Special Committee, which undoubtedly gave an upward bias to the number of times that subject came up. The bulk of the responses fell into a finite number of categories.

Issue 1: Stability and maintenance of the core infrastructure (hardware, software, human capital)

Out of the 40+ conversations, some variation on this general theme came to the forefront 23 times, more than any other topic. The cognoscenti widely recognise that demand for OSM’s services (tiles, geocoding, and geodata) has outstripped the current configuration of hardware and software, and is straining the volunteer labor force, particularly the sysadmins. They also widely recognise that the Operations Working Group has collapsed, which they see as symptomatic of the increased demands placed on the hitherto “100% volunteer do-ocracy”.

Though not universally held, the view that OSMF should begin to hire staff to augment the volunteer labor (particularly sysadmins and a software developer to maintain the API) is widespread. Notably, individuals responsible for maintaining OSM infrastructure specifically expressed that sentiment.

Several respondents called for the Board to act on the slow tile service, though proposed solutions ranged from “OSMF should invest in enough tile servers to meet demand” to “OSMF should restrict tile use to OSM only and let commercial providers do the rest.”

A few people advocated movement to the cloud. Lack of redundancy of some critical services was a concern to people familiar with OSM’s hardware setup.

Multiple old-timers made comments to the effect that “nobody envisioned OSM’s success,” or as one of them put it, “OSM shouldn’t have worked, but somehow it did, with lots of time, effort, emotion, and pride.” One respondent commented that OSM serves up maps to two billion users per day, yet has only two sysadmins to maintain and service the hardware. Several remarked that OSM has outgrown the model of “do-ocracy” that got it this far and opined that it needs a new management model. One respondent highlighted the “disconnect between casual mapping and the size of the project” as it is today, and noted that having grown from a “smaller project, and not especially geographically diverse” OSM today requires greater commitment. Another said, “More structure and more governance is [sic] required just because it’s used more extensively.” Put another way, a respondent said, “Growth and success have led it past the type of individuals who started it.”

Countering that, one respondent commented, “The Catch-22 of OSM is that actual mappers want a smaller OSMF” and don’t want dependence on “outside money.” One respondent perhaps captured this dilemma best: “Super organization isn’t necessary, but anarchy is not an answer, either.”

(Second place tie) Outreach to Local Chapters and Communities

Local chapter/community outreach tied for the second-most frequent issue raised in the calls, with 12 mentions (though only 9 local chapters/communities were contacted). These respondents commented on the desirability of better communication between the Board and local communities.

(Second place tie) Vector tiles

Surprisingly, a desire for vector tiles came up 12 times in conversations, tying for second place with community outreach. Some respondents merely see in vector tiles a sign of progress, that OSM is keeping up with Google; others see in them a solution to desires (multilingual standard maps, for example). While the tech wizards asserted that OSMF hardware is adequate to host vector tiles, at least initially, surprisingly not one respondent could quote a solid cost estimate for shifting from raster to vector tiles, nor could any respondent chart in shorthand a course of action needed to carry out such a shift. Time estimates ranged from “a couple of weekends” to “six to eight weeks”. One major issue appears to be who would control the style sheets that determine which vector maps are displayed, and some users in the commercial sphere expressed concerns that OSMF-sponsored vector tiles would compete with their paid services. In short, though there is a strong indication that vector tiles are desirable, there is lack of consensus either as to how much would be too much or how much it would cost.

One respondent noted that there is nothing to stop local chapters or others from hosting vector tiles if they wanted to, and suggested encouraging local communities and chapters to experiment with vector tiles before the OSMF settles on a solution, should it decide to do so.

(Fourth place tie) Frustration with Board inaction

Out of over 40 conversations, “Board inaction” tied for fourth place (with diversity/inclusion, at 10 responses) among the most-mentioned issues the communities wanted the Board to deal with. Respondents widely view the Board as having failed to take responsibility for issues that have arisen. One respondent asserted that the Board has taken exactly one significant decision since 2010, the change of license to ODbL.

One consequence of this is that third parties unaccountable to the community at large have filled some vacuums. The Board’s conscious decision to take a hands-off approach to development of the iD editor, in particular, is a flashpoint in this regard. While some welcomed development of a user-friendly, intuitive editor, even if by a third party not under community influence, iD’s tagging presets have raised concerns about perceived lack of community input into development decisions. As one respondent put it, “Key technologies should be OSMF’s responsibility.”

Another respondent viewed this abdication of responsibility as a prelude to long-term death of the OSM community, as it paves the way for a backdoor corporate takeover of OSM. Another respondent said bluntly that Board weakness “creates a power dynamic with outsiders who can pay workers–and then control it.” One respondent noted, “There is room for the Board to be more assertive because of the threats out there–the need to meet threats and challenges. Volunteers cannot do it themselves.”

Were the Board to begin making substantive decisions, in the opinion of some respondents another weakness would quickly become apparent. As one put it, the Board “has no real ability to put contracts in place” to implement decisions. The Board, this person said, “must either build capacity” or “let outsiders do it.”

Another issue is the Board’s failure to enforce its policies, such as tile use policy, which has led directly to a massive overload of OSM’s tile servers (which according to the Operations Working Group at peak loads respond to nearly 400,000 tile calls per second). Respondents raised protection of the trademark a few times, and thought that previous boards had neglected it. Additional issues include failure to pursue community development, a bias in favor of European points of view, and failure to demand attribution for use of OSM data.

However, two prominent community members asserted that Board inaction is “the OSM way” and indicated a desire to continue to see the OSMF Board as a mere figurehead, existing solely to fulfill requirements of the Companies Act 2006 and nothing more. As one put it, “The Board is to do the minimum necessary to keep OSM running.” Another expressed fear that a future Board could “go in a bad direction” and thus the precedent of the Board’s making decisions “could bode ill.” The community will need to reconcile this divergence in attitudes one way or the other.

(Fourth place tie) Diversity/Inclusion

This tied for fourth place with “Board inaction” (again, with probable upward bias in response rate since the chair solicited feedback on relevant Board actions). Communities outside western Europe generally welcomed the Board’s recent adoption of a diversity statement and formation of the Diversity and Inclusion Special Committee (with the caveat that members of one local community had not read the statement and so could not comment on it). One community went so far as to say it was long overdue, but still is not enough because local communities avoid speaking up out of fear of very vocal and hostile community members in other geographic areas shouting them down. One respondent said bluntly that the Diversity and Inclusion Special Committee “needs a space for discussions without being attacked” and cited “a tendency to intimidate” on the part of other community members. In that regard, multiple respondents called for a Code of Conduct of some sort to moderate dialogues and reduce the fear of hostile responses.

Respondents in Africa and Asia underscored the cost of volunteering, noting that in lower-income countries the cost of Internet access and need to work more than one job to support a family constrains time devoted to volunteer mapping. This is an obstacle to geographic diversity.

Respondents see the fee waiver program for Foundation membership positively in theory, but in many minds, its impact remains to be seen. Data users were surprisingly supportive of diversity because they see it as a source of data quality. As one respondent put it, “mapping is somewhat subjective” so diverse mappers generate more diverse (i.e., complete) data than does a “white male dominated” mapping community.

Many respondents complained of special interests’ “steering” or “dominating” issues to the detriment of the broader interests of the OSM community. As one of them put it, “If you let the loudmouths direct strategy, nothing will happen.” Another put it slightly differently, “Use of the project is imperiled by a few loud voices.”

One respondent in Europe criticized the Board for focusing on “political correctness” by publishing the diversity policy and forming the special committee.

(Sixth place) Artificial intelligence/machine learning

The sixth-most raised issues revolved around artificial intelligence and machine learning, with those in favor of incorporating them under human approval processes represented in Asia, Africa, and among the corporate members sponsoring these technologies. Opposition to artificial intelligence and machine learning seems to be concentrated in western Europe, where it is viewed as of little utility. Support is found in geographic locales facing daunting obstacles: high internet costs, low internet penetration, and low volunteerism, the latter two often rooted in economic circumstances. As one respondent put it bluntly, “[our country] is vast, and we don’t have enough volunteers to map all the roads and waterways by hand.” Mappers in such circumstances appear to welcome AI tools as a way of increasing craft mapper productivity.

Interestingly, the corporate members underscored continued importance of local knowledge, for as one put it, “AI can draw a road, but only a local mapper can name it” (and we might add, certify that it is a road and not something else.) Another non-corporate respondent noted, “A growing proportion of data cannot be collected by armchair mapping, we need on-the-ground knowledge.” Corporate users, who professed keen interest in improved data quality, highlighted the role of AI in rapidly detecting vandalism so that the Data Working Group and local mappers can react quickly.

While application of AI and machine learning is not a Board issue, strictly speaking, but rather one squarely in the laps of the local communities each to decide within the framework of organized mapping guidelines, its importance to certain local communities was striking.

Comment from imagico on 18 April 2020 at 17:49

Very interesting read, thanks.

While this is what i would classify as anecdotal observations of course it is a large volume of them and as such represents an interesting cross section of views from within the community.

Two observations about that

  • It should be clear that this cross section is of course in no way representative for either the OSM community as a whole or the active OSMF members or some similar group of people. There is subjectivity in the selection of the people you talked to as well as selectivity in your perception of the conversation. That does not in any way devalue the communication and the ideas and opinions communicated but quantitative assessments in particular - like the ordering of priorities or statements like “Communities outside western Europe generally welcomed…” are of course to be interpreted with that in mind.
  • While possibly not everyone will agree with that i think that decisions on matters like the ones you listed in the Top Lines require looking at the merits of the ideas presented and should not just be based on their popularity. In particular it is paramount to think through even very popular ideas beyond the point to which they have been considered by their proponents. For example the old story of “We live in a vast country and just don’t have enough volunteers so we need imports/automated AI mapping/paid mappers/…” is rarely thought through up to the point when it comes to the long term maintenance and updating of the information over decades by those who tell it.

On quite a few of your Top Lines by the way people (yes, including me - but this is not about me) have written down stuff on blogs, diaries and elsewhere that probably contains a lot of meaningful considerations much of which has likely not been in your records of conversations. In oral conversations - and also often in people’s recollection of such conversations - a fairly large focus is often put on what people want and if they are for or against something. But often IMO the more interesting part is the motivation and reasoning why people are in favor of or against a certain idea. This in my experience often becomes clearer when people invest the time and energy to formulate their ideas in writing.

Comment from migurski on 18 April 2020 at 19:17

Allan, thanks for starting your time on the board with this sustained effort to reach distant corners of the community. It’s inspiring to see you take your chairmanship responsibility so seriously and to gather so much input beyond just the visible contributions.

This point about obstacles to joining a “do-o-cracy” is especially significant and I’d love to have seen it emerge as a more widely-shared concern:

Respondents in Africa and Asia underscored the cost of volunteering, noting that in lower-income countries the cost of Internet access and need to work more than one job to support a family constrains time devoted to volunteer mapping. This is an obstacle to geographic diversity.

Comment from Peda on 18 April 2020 at 20:28

Hi Allan,

interesting read, thanks for collecting and your efforts!

I think it is debatable if you could/should state who you spoke with, as the viewpoints are aggregated and only contain unattributed quotes. Still if you (or the persons interviewed) demand to be unnamed this has to be accepted. I am still interested in a full list though :-)

Anyway, I’d at least be very interested in some basic information of “statistics”. I share many (or at least some) of the points mentioned in your summary. But somehow I’d like to also understand in how far this survey is ‘representative’. Key facts I’d be very much interested in:

  • How many of these conversations were with people from the US, Europe, Africa, Asia?

  • How many of the participants were male, female, diverse?

  • How many of these conversations were with people having a commercial background compared to the ones without? (e.g. the AI section sounds like you talked to all corporate member)

  • To how many corporate members did you talk to? In how many different conversations?

These would be the questions I am interested in. While I understand that you probably didn’t specifically ask anyone for e.g. their gender, a rough estimate should still be feasible.

Thanks,
Peda

Comment from apm-wa on 18 April 2020 at 21:16

Peda,

Since the conversations were confidential and pursuit of honest views is more important than full disclosure of identities I will not divulge with whom the conversations were held (42 years of holding a government security clearance has taught me the value of holding at least some information in confidence). However, I believe I can answer some of your questions on statistics without revealing identities:

  • 43 conversations to date
  • 6 individual conversations with females ** females were also present in two group conversations, but I didn’t keep a tally so cannot tell you how many were in either group
  • 9 LCC conversations: 2 Africa, 3 Asia, 3 Europe (not counting a second conversation with one of them with followup questions), 1 North America
  • 9 corporations/private firms in 11 conversations ** some conversations were with people who are employed by companies using OSM data but who spoke from personal viewpoints and not on behalf of their employers (this is another reason behind confidentiality)
  • 5 of the conversations with LCCs were with people of color (Africa and Asia). I don’t know if anybody else I talked to considers her/himself a person of color–I didn’t ask.

If I were to give you more detail than that, you could probably deduce some identities, so I will stop.

If you have suggestions for people I should hear from, please provide me via e-mail with cyberintroductions to anyone you think would be worth asking the two questions I have been posing. Many thanks.

apm

Comment from jgc on 19 April 2020 at 08:57

Thank you very much for this important Outreach, and this report on it.

For the sake of Diversity and Inclusion, I have had it translated into French (with Deepl.com), and posted on a wiki page, so that the translation can be further reviewed and improved : https://wiki.openstreetmap.org/wiki/User:Jgc/TraductionJournalConsultationsApm-wa

Best regards,

Jean-Guilhem

Comment from RobJN on 19 April 2020 at 10:48

Thanks Allan,

I think this is the most extensive attempt any board member has taken to gather input from the community ever. Possibly more extensive than any non board member too.

I am therefore not sure I understand some of the concerns in the responses here, perhaps because they lack a way forward.

The way I see it is that the board is elected so are free to get on and do something. They are free to gather information and evidence however they like from informal chats, to “requests for information” to a Resolution of the members (as per the articles of association). At the end of the year we can pass judgement on this via the board elections.

If you’re looking for a more complex arrangement then it would have to be a tiered/hierarchical arrangement with defined groups and decision making powers. However so much is missing here at the moment.

Start at board level and let’s grow out from there until we find our suitable solution that goes beyond the previous “do-ocracy” but still meets the spirit of our project (i.e. not a super sized organisation).

Best regards Rob

Comment from apm-wa on 19 April 2020 at 11:58

Rob,

Thanks for this. I didn’t sense from any of the respondents a desire to grow into a “super sized organisation” as you gently put it. Rather, I sensed pride that OSM does so much with so little, but while acknowledging that, a recognition that the infrastructure is at risk so something needs to be done to assure a stable platform going forward.

In a sense we already have a tiered/hierarchical arrangements with the working groups. Two-tiered, to be precise: Board->working groups. So far in the conversations I did not sense a desire from any quarter for that to change, or a particular need at this point.

As an aside, I should share with you my own analysis of why OSM has not evolved along the same lines as Wikimedia, i.e., why Steve Coast is not a clone of Jimmy Wales (apologies in advance to Steve). As an economist among other professional qualifications I periodically put on my econ spectacles and view the world through that prism.

  • Wikimedia offers a free good to the public but has no way of monetizing it once it is published, thus for Jimmy Wales to support himself and his avocation, he needed to grow his NGO into something attracting large enough donations that it could provide him and his colleagues livable salaries.
  • OSM offers a free good to the public but as raw data that can be transformed into a higher-value product, which in turn may be sold, which is why a commercial ecosystem evolved in parallel to OSM.

That commercial ecosystem is the source of livable salaries for much of the OSM community, as well as the corporate donations. These are two superficially similar volunteer-oriented models with a deep and decisive difference, and that difference is why OSM is unlikely to grow into something like Wikimedia. It simply doesn’t have to. We will find our own way, and most probably remain relatively small.

apm

Comment from Tordanik on 19 April 2020 at 12:52

Rob, be assured that members of the board spend thousands of hours on the community’s communication channels, at OSM events, and in direct conversations with other members of the community. This may not be explicitly labelled “outreach”, and we aren’t usually writing blog posts about it because it’s just a normal part of daily life for us. Still, it serves many of the same purposes of keeping an ear to the community’s concerns and informing our next steps on the board.

Allan’s outreach strategy complements our usual methods in that it reaches some people who we otherwise don’t hear from as much (and vice versa, of course). Our 2019 Pre-F2F and Pre-SotM surveys had similar benefits in unlocking input from additional corners of the community.

But please don’t assume that we don’t listen as a matter of course, and that this is more than the tip of the iceberg. If any of the community members who read this have something to share: You don’t have to wait for us to call you! Your views are valuable, so please don’t hesitate to discuss them publicly on one of OSM’s platforms, or to approach the board or one of its members directly if you prefer.

Comment from RobJN on 19 April 2020 at 15:12

Hi Tordanik,

I know. But in my experience there is a difference between passively listening and actively engaging. I certainly appreciated Allan’s approach and hence my comments. This wasn’t meant to belittle the work of others so sorry if it came across that way.

As you know our community is large and complex. It simply will never be possible to read everything in every channel. Indeed I don’t want the board to be doing this as it will leave no time for anything else. You also wouldn’t get a full picture as on some channels the views of the vocal minority don’t match the views of the silent majority.

I think the right solution is to get the info you need at the time you need it related to a policy or change you are working on.

Comment from RobJN on 19 April 2020 at 15:24

Hi Allan,

In a sense we already have a tiered/hierarchical arrangements with the working groups. Two-tiered, to be precise: Board->working groups. So far in the conversations I did not sense a desire from any quarter for that to change, or a particular need at this point.

Well indeed, but there is some desire to refine it a little. As you said yourself many “widely recognise that the Operations Working Group has collapsed”. On the LCCWG (on which I am a member) we have more on our plate that we can currently achieve.

So yes, there is a structure and no radical change required. However in my view there is presently limited ability for the board to delegate as much as a board normally would due to limitations in the WGs. Some can be solved by establishing new groups (DISC, microgrant committee) others will take a little longer.

Anyway I don’t think we are disagreeing here, a least not fundamentally so I’ll stop talking as I seem to have been misunderstood by both you and Tordanik. Perhaps another sign of why video calls are so important. :-)

All the best, Rob

Comment from apm-wa on 19 April 2020 at 15:49

Hi, Rob,

In addition to the limitations of numbers of WGs we have a limitation on the number of volunteers currently engaged in the WGs, and that is a serious constraint. We need to recruit more people to assist via the WGs, and any ideas you and others have on how to do that would be welcome. The DWG and OWG are particularly hard pressed.

No, I don’t think we disagree seriously.

Regarding your comment, “You also wouldn’t get a full picture as on some channels the views of the vocal minority don’t match the views of the silent majority,” you are correct that part of my motivation (and part of the reason some of the individuals and groups I called were pointed out to me by others) was to hear from people who do not routinely publish their opinions. It is true that we receive a great deal of unsolicited advice on how OSMF should function, but I believe that active outreach is an important augmentation to that flow of information.

apm

Comment from RobJN on 19 April 2020 at 20:18

Unfortunately I don’t know much about either the DWG or the OWG. I’ve usually been better at the “softer” skill set, hence why I was on the SotM WG for 2013, and 2015 to 2018. I left that group mainly due to a lack of time, but also because there were some members of the OSM community that would always complain no matter what you did. After so many years staying up past midnight to support SotM during busy periods (including on the night before my paid work’s biggest day in the annual calendar) it had taken it’s toll on me. No idea if DWG or OWG members feel the same.

You mentioned above that some of the corporate members “highlighted the role of AI in rapidly detecting vandalism”. Perhaps the DWG could look there for extra support? Sounds like they all want the same outcome (best local data for improving OSM).

Other than that all I can offer as a suggestion is a targetted promotion campaign. We have done mentions at Foundation election time and during SotM. Sounds like we need something more targetted or sustained.

Failing that, how about paying for what we need? It sounds like there are a lot of people without work now due to Covid-19. Perhaps it is a great time for OSMF to step in with some short term contract work. Gets us what we need and helps a few people during what is clearly a tough time globally.

Comment from SimonPoole on 19 April 2020 at 21:16

Hi Allan

Wrt the similarities and differences between Wikipedia and OSM, I do think it is important to note that it is not just a question of differences in distribution of the product. WP has also enjoyed a near monopoly for a large part of its history, a luxurious situation that OSM has never had.

There are essentially no replacement products and vendors for what WP does (because they were killed off nearly immediately), and lots, if not all, of WPs marketing (for donations) revolves around that they are not replaceable. So I would take a bit of an issue with the statement that they don’t monetize their content.

Very different with OSM, there are plenty of players at all kind of levels, from local to global, that are offering viable replacements and alternatives.

“out-sourcing” end-customer services and products, has out-sourced the financial investment and risk for the OpenStreetMap project too and allows us to operate without directly competing with all the other players.

Simon

Comment from apm-wa on 19 April 2020 at 21:53

Simon, I use the term montization in the sense found in the Wikipedia article “Monetization”, section “Revenue from business operations”. ;-)

Comment from SimonPoole on 19 April 2020 at 21:57

Well historically the WMF has offered bulk APIs for some “customers” :-), I would have to dig up the financials on that to see if they directly sold them or not. But my point would be more if you have a monopoly for something very in demand you don’t need to “sell” it in conventional sense, you can just threaten you will stop providing it if people don’t donate.

Simon

Login to leave a comment