RESPONSES_TO_RFI_QUESTIONS.pdf

Disclaimer: GAO is still planning the requirement so these answers should not be considered to be part of any future solicitation.

Question/Comment

Please provide the total number of taxonomies, the average number of terms per taxonomy, largest taxonomy size and total number of terms across all taxonomies.

• There is currently one taxonomy currently being used. There are approximately 9,000 terms currently in use in it.

What is the expected rate of growth of the taxonomies, once the new system is in place?

• This is undetermined beyond the plan of adding more taxonomies once the system is in place.

How many systems will be requiring "database level access" to the taxonomy?

• At least three systems will need ‘database level access’ to taxonomies, with others possibly added in the future.

Please provide more information on the types of externally-facing applications that will need this level of access.

• GAO’s Drupal environment and the internal Product Metadata Repository will are applications that will need this level of access. Other applications may need access at a later time.

Will these externally-facing systems allow users to search the taxonomy? If so, how many queries are expected at a) average and b) peak times?

• Yes. See the metrics below.

• Time period: Mar 7, 2021 to Mar 5, 2022

• Total Sessions without search: 5,578,681 (0.37% of visits used search)

• Total sessions with search: 20,877

• Total unique searches: 21,357

Public Website Usage Info (3/7/2021-3/5/2022)

Daily Usage Users Sessions Pageviews Sessions with Search

Average 13,849 15,383 29,615 57

Minimum 4,924 5,252 7,731 11

Maximum 34,300 37,343 5,916 181

How many documents total will require auto-classification? How many documents are added/changed to the document repository per day? How much is the document corpus expected to grow over the next years?

• GAO releases 600-800 products per year, which averages to about three per business day. The document corpus will grow as GAO continues to release products over the years.

What document types will require auto-classification? Where do these documents come from?

• The most common document needed auto-classification will be GAO products in PDF form.

HTML files, Word documents, and plain text files will also likely be needed. The documents currently come from the publishing group in the IT department.

Will the classifications be used to make search available to external users? If so, what is the expected search volume?

• Yes. The search volume is in the table above.

How many staff persons are currently involved in writing classification rules? Is this a full-time activity for them?

• There are currently only two staff that have worked on writing rules for the legacy system. This is not a full-time activity currently, as they are not being edited right now.

What level of FedRamp does GAO require (Moderate, High, etc.)?

• Low, with a slight chance of moderate if any PII is included.

How large is the GAO Taxonomy team that will require training?

• Two people will need training for using the system. Others may need training for integrating the system into IT applications.

Please can you provide copies of the relevant Program Orders and Records Management Schedules.

• Yes. Records Management Order (0410.1) can be provided. Administrative records retention is 7-years.

Please clarify this requirements. Are you expecting the vendor to develop familiarity with the legacy system?

• No. We would just need a well-documented API that could be integrated internally.

Is this a contractual requirement to complete the implementation in 3 months?

• This sounds about right from the vendor perspective of configuring and giving access to the system, with an understanding that an ATO could take at least 90-days. Internally we’re guessing 6-8 months of PMR/GAO.GOV integration.

1. Reference: Synopsis, p. 2

• Comment: This Section states, the contractor “shall offer an IT solution that is compliant with the legislation’s Agency Program requirements.” It is unclear what the specific requirements are.

• Question: Will the GAO please identify the specific Agency Program requirements that are being referenced?

• Please see pages 3-9 for the requirements being referenced.

2. Reference: Transition to New System, p. 6

• Comment: This Section addresses services to migrate from its current taxonomy server. However, there is no information on the data to be migrated.

• Question: Will GAO please provide a description of the data that is to be migrated – volume, data types, etc.

• An Excel file detailing the structure of the revised GAO Product Taxonomy will need to be ingested into the new system. Also, text (.txt) files of previous versions of the taxonomy may need to be ingested.

3. Reference: Required System Components and General Functions, p. 4

• Comment: The section states “A.8.1 The system must allow creation of at least five (5) user accounts that allow for the permission levels below”.

• Question: Does the customer desire user accounts to be unique to the Taxonomy Server, or should they be imported from a managed authentication server (Microsoft Active Directory, or other)?

• They should be unique to the Taxonomy Server.

4. Reference: Required System Components and General Functions, p.8

• Comment: The section states “C.2 The vendor shall regularly enhance the products based on customer input and industry changes.” The scope of the enhancements is unclear.

• Question: What is the expected scope and frequency of the enhancements?

• Enhancements will be made to address errors (both technical and interface related) in the system and to align with updated W3C standards. The timing of these changes is undetermined, as necessary changes cannot be fully anticipated.

5. Reference: Required System Components and General Functions, p.8

• Comment: The Section states “C.4 The vendor shall provide all updates to the system without further costs to the agency.” The scope of the updates is unclear.

• Question: Can you define the scope of the updates?

i. Security updates and other routine software updates are included.

6. Reference: System Architecture and Security, Page 6 B.

• Comment: The COTS taxonomy solution to be proposed requires a repository to store data. The solution has already been integrated with the Knowvation CSP under a different government contract that is owned by the library and is imbedded in the LibLime Bibliovation LSP software currently used by GAO. If we could use that solution, it would provide cost savings to the government while also allowing easy Taxonomy integration with the other library information systems, should that be desirable in the future.

• Question: Is it acceptable to leverage that investment and propose to use that repository for the Taxonomy solution?

i. No, it is not acceptable.

Reference

(Document name, Section #, page #)

Question/Comment

1 General Which vendor provided support on the previous contract and when did this contract end?

SAS is the legacy system provider. The contract is currently limited to a maintenance contract; no taxonomy editing support is being provided.

2 Page 3, 1st paragraph How many levels deep does the new taxonomy that is being designed go?

Three levels.

3 Page 3, 1st paragraph Can the GAO provide the draft taxonomy as an attachment to the final

RFP?

No.

4 A.1.2

Is the vendor expected to integrate the GAO systems with the software, or will the GAO staff undertake this work?

No. GAO staff will undertake the work.

5 B.3 If the data center is cloud-based and split across multiple locations, what level of access does GAO need?

The system must allow permissions assignment at the term level. Access should allow at least three of these permission levels: Read- only comment only, editor, and system admin (control for all system permissions).

6 D.1 How much data (in GB, TB, PB) will be scanned from the GAO’s system?

The agency has roughly 1200 products a year, close to 189k total, and a recommendation count of around 67k. This is slightly over 5GB of data.

7 General The RFI indicates no updates of the current server have been done for 3 years due to lack of product support from the vendor. Please clarify if there will be any vendor support available for the migration of data from the current system to the new system.

There will not be legacy vendor support for migration.

8 General The RFI indicates no updates of the current server have been done for 3 years due to lack of product support from the vendor. Does the GAO mean the SAS (company), or a SAS integrator was not able to provide support?

SAS does not offer a more recent version of the software currently running. The contract in place is limited to maintenance.

9 General Are multilingual capabilities desired/required?

Multilingual capabilities are not required.

10 Section VI. Delivery Please provide more details on training requirements? Specifically,

- number of users/role,

- in person or virtual,

- train the trainer or end-user,

- training materials needed or not, etc.

• Training will be for two people, who will be the most advanced users, at an administrative level.

• Virtual training is required.

• Those to be trained will be administrators and advanced users of the platform.

• Documentation of how the system is used as an administrator and advanced user is required. If that is not provided by already existing system documentation, training materials are required.

11 Page 2, Background In the Background section on page 2, it is stated:

“The Center for Library Science requires a commercial off-the-shelf (COTS) taxonomy system. This requirement also includes the need for the migration of data into the new system and then on-going support and maintenance of the COTS system.”

Is the data referred to above to be migrated, the taxonomy metadata associated with each GAO document in the archive, or is it also referring to the GAO source document archive data as well?

The data referred to is the taxonomy metadata associated with each GAO document in the archive.