An online Unidata Survey 2001 was developed by the Unidata Users Committee to solicit responses from two key groups:
The Unidata Survey 2001 was distributed by email on three separate occasions between April and October 2001 using Unidata’s all-community distribution list. This list includes approximately 1100 addresses with some duplicates and several non-university contacts. A total of 148 surveys were completed and returned to Unidata: 68 from Organizational Representatives and 80 from Individual Users. The majority of responses were from universities using Unidata data and applications, however, approximately 10% of the responses came from non-university organizations such as NOAA, USGS, and Woods Hole Oceanographic Institute.
The Survey sought input on a variety of issues ranging from current needs to future expectations regarding hardware, software, and data, system management, the relevance and importance of emerging Unidata initiatives, disciplinary needs and trends, and overall satisfaction with the Unidata Program Center and Unidata Governing Committees.
The 2001 Survey was the first community-wide assessment since 1993, and the second since Unidata’s inception. The purpose of the Survey was to solicit input directly from the community of users that could be used to provide guidance to the Unidata Program Center and Unidata Governing Committees for the allocation of resources and improvements to the quality of service, while serving as an assessment of the level of interest in and relevance of emerging Unidata initiatives. Insights gleaned from the Survey are being incorporated into Unidata’s Strategic Plan and the Unidata proposal to NSF, and will be used for internal decision making purposes.
The Unidata Survey 2001 revealed few surprises. This can be interpreted in two ways, both of which are in part correct. First, the Unidata Program Center (UPC) seemingly has its “finger on the pulse” of its community of users to a degree that there are no areas of service and support where a significant disconnect exists between the UPC and the needs of its community. Second, the initiatives being spearheaded by UPC in cooperation with the community have obvious benefits to users and are accepted by the community as natural and logical progressions by an acknowledged leader in the areas of data access, software development, and support services. The survey portrays a broad and diverse community of users with meteorology/atmospheric sciences as the core constituency comprising about one-third of an extended community that includes Climate/Global Change, Oceanography, Geography/GIS, Hydrology, Earth System Science, Environmental Science, Geology/Geophysics, and Computer Applications. By simply completing the survey and indicating their awareness of Unidata programs and mission, it can be inferred that current initiatives on the part of UPC to collaborate with these non-core scientific arenas are at least partially successful. In the spirit of Unidata interactions with various communities, these collaborations are distinctly multi-faceted and multi-directional, in that it is just as likely for an organization/user to seek out Unidata to help accommodate some need as it is for Unidata to entrain an organization/user in order to make available a service for the greater good of the community-at-large.
The survey provides UPC with clear mandate to proceed responsibly with the further development of emerging initiatives such as MetApps, THREDDS, platform independent software, smarter push-pull LDM software, and a Web-based portal (PlazaElectra). A highlight of this survey is the 97% approval rating for the UPC, and satisfaction rating of over 80% in the categories ease of use, education value, research value, and technical support. There are few organizations that can achieve this high level of approval for its services, especially while serving such a diverse community of users. These ratings are testimony to the sustained commitment and foresight of the UPC management and staff.
An interesting caveat evident from the survey is that, while users are largely satisfied with the data, software and support provided by UPC, a majority of users are unaware that they are being directly represented by three governing committees: Policy Committee, Users Committee, and Technical Advisory Committee. Unidata is a community-governed, community-driven program, yet the governing committees and their members, selected so as to represent an approximate cross-section of the Unidata community, are largely transparent. Several interpretations of this survey result have been suggested, most of them with a slant toward this being a problem, yet given the results that indicate a high level of satisfaction with UPC and its services, which are largely an outgrowth of community-based needs and issues to which the committees respond and recommend action, it could be inferred that the committees are functioning in accordance with the wishes of the community.
If there is a problem endemic to the governing committees, it is the difficulty associated with the dissemination of information related to products and initiatives, new data types, software, actions and recommendations of the governing committees, and survey results to the Unidata community. Part of this problem is directly related to the Unidata Web page, a portal that is currently extremely unfriendly and difficult to navigate. In progress is a radical revision of the Unidata Web site, which should resolve many of these problems by creating an interactive portal, MyUnidata, that can bring information to the user, and request input from the user, within a customized Web-based framework.
The Unidata Users Committee conducted a detailed analysis of the Survey results in order to ascertain the needs of the community. The assessment contained herein is a summary of this analysis.
It is quite significant that the Unidata community continue to expand into non-traditional, yet related fields. Traditional Meteorology/Atmospheric Science now represents only a third of the disciplines using Unidata applications and data. While Meteorology/Atmospheric Science remains the core user group, and is not likely to relinquish its primary role in guiding the direction of Unidata in the foreseeable future, targeted outreach to related communities, while being mindful of resources, can only serve to stimulate and enhance the development of new analysis and visualization tools, and broaden the availability of and accessibility to various data sets, to further encourage interdisciplinary approaches to scientific, geographic, and societal problems of concern to the expanding Unidata community of users.
Two-thirds of individual users dedicate over 50% of their time to research (Question 3). It is not clear from the survey if the same researchers use Unidata applications and data during the time spent (£ 50%) in the classroom, but based on the numbers of students that are directly (5600 students) and indirectly (7700 students) exposed to Unidata applications and data, one can infer that Unidata applications and data are seeing substantial usage in both research and education. This inference is strengthened by the organizational response that, of the total number of academic and research staff (2703) cumulatively represented by the organizational survey, 69% make some use of Unidata applications and data for research and education. The fact that the majority of students exposed to direct (hands-on) use of Unidata applications and data are undergraduate non-majors, with healthy representation from the undergraduate majors and outreach activities sectors, suggests that Java-based, platform-independent applications such as MetApps, are likely to see increased student usage for display and analysis of real-time and archival data sets.
As Unidata looks to the future, guided by its Strategic Plan, it is imperative that it remains cognizant of the number of institutions that continue to rely on faculty members and/or students for computing systems management. Over 45% of the institutions responding to the survey claim that organizational support for computing systems is sustained by academic and/or research faculty or undergraduate and graduate students. More than likely these are smaller to moderate size schools with tight budgets that use Unidata applications and data for classroom instruction and operational forecasting. Unidata cannot lose sight of the importance of easy downloads and installs, seamless data transfer, intuitiveness of applications functionality, documentation, and support for this community of users.
Unidata made a decision a few years ago to incorporate Java-based, platform-independence into development of its applications. MetApps is a prime example of Unidata’s commitment to this programmatic strategy. Organizational responses indicate that a parallel and concurrent move toward platform-independent systems running Linux and Windows is occurring at universities.
The use of McIDAS was surprising in light of discussions among members of the Unidata Users Committee that generally assume that more users employ GEMPAK. The survey does not bear this out. Moreover, with the newly developed high-resolution national radar composite, which must be viewed using the McIDAS GUI, and with ADDE architecture being incorporated into MetApps as a remote file transfer protocol, McIDAS is well poised for sustained development. GEMPAK continues to be a workhorse application used by many institutions for education and research. Many of the features of both packages are being incorporated into MetApps functionality.
Over half of both organizational respondents and individual users indicated the FX-Linux/FX-Net was worth investigating but not at the expense of shifting resources. Considering that another 31% indicated that it was unimportant, the resounding message from the community is to not direct resources toward this activity and away from other important applications development. Universities apparently not interested in replicating NWS computing systems for education and research if the result is a reduction in the development and support of other applications they perceive as more important to their needs. It could be that the community does not want to see Unidata spread too thin. Another interpretation is that universities do not see themselves as a computing systems training ground for NWS meteorologists. Rather, universities remain committed to analysis and visualization applications that can convey straightforward concepts using architecture that can function efficiency in both a research and classroom setting.
The widespread usage of non-Unidata applications by the community, many of them commercial, should be considered when building new applications with transparent interfaces and common data formats.
Organizations and users are employing Unidata’s collaborative resources: COMET, SuomiNet, and CONDUIT. The highly successful Web-based COMET modules are used extensively in educational settings and are likely to see continued usage as new releases come available. Unidata’s IDD is used to relay high resolution GPS to UNAVCO (GST), where solutions are derived to produce 2-D products of precipitable water vapor and total electron content, which are then distributed via the Web. Only about 30% of the 100+ registered SuomiNet sites are currently transferring GPS data. As more universities come online, the volume of data will increase. In addition, universities are encouraged to produce 3-D products for dissemination to the community, and numerical mesoscale modelers are already experimenting with data assimilation methods to incorporate SuomiNet data. CONDUIT has not made significant inroads into the university community except in select instances were an individual user is interested in archival model data. However, recently the Unidata User Committee selected a liaison to serve on the CONDUIT steering committee with the anticipated outcome being greater accessibility to CONDUIT resources.
The size and type of data sets is likely to experience significant increases during the period covered by the five-year proposal and the strategic plan. Level II radar data, ACARS, and other real-time data streams, currently being received in high volumes by only a limited number of institutions for data assimilation into mesoscale numerical models, is expected to increase dramatically if any one of several proposals are funded. Unidata has begun to address these data transfer issues by exploring the NNTP (Network News Transfer Protocols) as a potentially viable alternative to the push-technology of current LDMs. It is imperative that institutions have the ability to select and pull data from any number of distributed and redundant data repositories. In addition, as Unidata bridges to communities outside the meteorology-atmospheric sciences domain, data formats, metadata, and data types are sure to increase. Unidata’s mission is to provide data, applications, leadership, and support to a community of users. Historically that community has been institutions offering degree programs in meteorology/atmospheric science. Unidata must remain true to this community, while trying to find resources to develop inroads into other communities. The wide range of data types bodes well for applications like MetApps and THREDDS. The increase in data types and volumes is seen as one of the principal challenges to Unidata over the duration of the proposal five-year plan.
MetApps, THREDDS, and Plaza-Electra are emerging initiatives central to the five-year strategic plan and the proposal to NSF. Although too many in the Unidata community still do not know about these initiatives, the survey makes it apparent that respondents see value in each. MetApps, a community-built, next-generation interactive meteorological Java-based, platform-independent application has features that should quickly entice researchers and educators to begin moving away from platform-specific and more static applications such as Gempak and McIDAS when it is released. While some survey reviewers find it troubling that MetApps is not yet well known, it appears reasonable to assume that this will continue to be the case until this application emerges from its beta test mode. Researchers and educators just do not have time to “play” with something that is in the process of changing. Unidata workshops offered at both the UPC in Boulder and regionally would be an effective way of disseminating MetApps to the community. Unidata might consider previewing MetApps at the UCAR members meeting and the AMS/UCAR Heads and Chairs meeting in October.
The value and potential benefit inherent in the THREDDS initiative are self-evident as indicated by the number of respondents that selected somewhat valuable or very valuable as their response. THREDDS is essential if Unidata is to expand to meet the needs, and provide data, applications, and support to an ever-expanding interdisciplinary community of users. The cornerstone of Unidata’s success is the rapid delivery of data using a distributed system of servers. THREDDS will build on the experience base of Unidata’s highly successful IDD system to enable users to access real-time and historical data from a network of thematic servers seamlessly from the local desktop.
There are not many programs that engage such a large community of users who can boast a satisfaction rate of 97%. Unidata is unique in its service to the community, and its engagement of the community in application building and decision-making. It is a successful venture by any measure and provides a real benefit to the research and educational activities that take place at over 150 universities nationwide.
The large number of organizational contacts that feel part of the Unidata community is likely due to the fact that these individuals are deeply involved in maintaining the computing systems that ingest data and run applications, dialogue more frequently with Unidata support staff, and must deal with technical issues related to installations and upgrades that directly connect them with the Unidata staff. Users, on the other hand, are typically more passive participants not likely to give Unidata much consideration when the data feed is uninterrupted and the applications are working as expected. However, there is a caveat to this assessment, in that, individual users have less frequently attended User Workshops, maybe because they lack the time. But efforts should be made to reach beyond the organizational contacts to the users. The Unidata governing committees, in consultation with the Unidata staff and through suggestions made by users in this survey, should consider more effective ways of reaching the users and enhance their participation at Unidata workshops. The Users Committee should spearhead these efforts.
Is it that the Unidata governing committees are transparent that so many organizational contacts and users do not know much about them, or is it that the governing committees have not engaged the community effectively? The consequence is the same in that a large fraction of the community does not know if the governing committees are representing their needs. Efforts should be made to ameliorate this situation so that users are aware that they are being represented in the areas of policy and users needs. Based on the results of this survey, an ongoing effort on the part of the Unidata Policy and Users Committees in particular should emerge to make the community aware of committee activities, decisions, action items, deliberations, and outcomes. A significant upgrade to the Unidata web site would be one way of connecting with the community. The web site is in dire need of a major enhancement. A second, and more appealing way would be to proceed toward the development of personal profiles, “MyUnidata,” where committee highlights would be available to an individual that entered through a web portal.