South african government procurement

The problems with GIS-related procurement in South African Government

• The market is dominated by expensive, proprietary, foreign-based vendors (like ESRI, Hexagon and Oracle), resulting in stifled competition and lack of opportunity for local economic development. Section 217 of the Constitution: "When an organ of state in the national, provincial or local sphere of government, or any other institution identified in national legislation, contracts for goods or services, it must do so in accordance with a system which is fair, equitable, transparent, competitive and cost-effective."
• Various Government policies and regulations are being flouted and have been for many years
• Procurement is generally very poorly executed, whether through incompetence, laziness, influence or poor advice
• The Governemt and hence the taxpayer is being extorted when far cheaper alternatives exist
• Service providers are often forced to purchase proprietary software to service by their government clients' bizarre, irregular, non standards compliant requirements

The situation has really become untenable (for everyone except the vendors) especially in this new environment of trying to clean up governance and reduce expenditure.

SAGI, CSI, SITA, GISSA, Practitioners, Competition commission? Why are we allowing this anticompetitive and unaffordable practice to continue, especially in light of the current economic and political climate in SA?

Our imperatives

• To force any Government entity whose proprietary ELA (enterprise licence agreement) is expiring to go out on tender for GIS and related software and services
• To force any biased or non-competitive tender for new or continued GIS or related service or software provision to be withdrawn, rewritten and re-issued

South African IT procurement, amongst many other factors and regulations, is governed by these principles

• Functionality
• Cost containment
• Open Source (FOSS)

An OPEN tender or quotation process must be followed in these circumstances

• when new requirements arise
• at the end of an ELA or SLA

These are the basic steps that ANY government entity is supposed to follow

• conduct a process to determine functional and non-functional requirements and write these up as formal ToR (Terms of Reference)
  • you may not mention a product or brand OR structure the ToR so they are biased towards a brand or vendor
  • a 'licence' is not a functional requirement and therefore has no place in a ToR
• put these ToR to OPEN tender or quotation (depending on size) via the CSD (Central Supplier Database) so they appear on the | eTender portal
• if a FOSS-based bid is submitted and is considered responsive, it has to be adjudicated fairly along with any other bid. If it comes out with the best score (technical+price+BEE) then naturally it must be selected.

It needs to be made clear to SCM heads, City Managers and anyone else in the decision chain that:

• they may not simply renew licences but have to go out to OPEN tender and consider alternatives
• they may not specify software by brand but have to specify a solution to a problem
• they cannot use the argument that ESRI or any other vendor is a sole supplier as there are many equivalent alternatives, both proprietary and FOSS
• Opportunity has to be given to FOSS solutions if offered since that is Government and SITA policy
• They may not use the ‘framework agreement’ to justify purchasing ESRI or Oracle
  • SAGI is pushing SITA, Treasury and the Auditor General for the framework agreement to be scrapped or in the interim for FOSS alternative to be added to it (which will leave any entity invoking the framework agreement with no choice but to use the FOSS option since its aim is cost containment)
• The ‘framework agreement’ in any case only applies to Provincial and National entities administered through SITA and is only a mechanism to secure discount pricing IF the continued use of ESRI or Oracle is justified (which it never actually is for Government)
• that the argument that ‘we’ve already made such a huge investment in ESRI and it will be wasted, or cost us a fortune, if we change’ is known as the sunk-cost fallacy and is exactly that, a non-argument. There is NO justification for hanging on to a proprietary system.
• that having invested in systems or processes that lock them in to a vendor such as ESRI is no excuse or reason for not changing. If they had complied with MIOS and other legislation and the Constitution itself, as well as best practice IT design, they would have ensured from the start that any system they invested in would be standards-compliant and interoperable, even if it was built with proprietary software.
• the use of the ’Section 36’ emergency provision is highly suspicious and will expose them to prosecution (deviation from SCM regulations e.g. https://www.iol.co.za/mercury/section-36-only-for-emergency-1219876u
• they are liable in any case for fruitless, wasteful and reckless expenditure in these times of austerity and widespread corruption and maladministration
• that perks like trips to international user conferences that are included in licence agreements and described as ’training’ could be contrued as kickbacks in disguise and that purchasing decisions can be skewed when these muddy the waters. Furthermore, I doubt the GIS users benefit from these as much as the managers who signed off on the purchase or insisted on the requirement.
• you may not procure a particular brand of software just because you or your staff like it or know it. You have to go through a systematic, rational, legislated process of finding the optimal solution for the problem at hand, whatever form it may take.

See some examples of tenders at http://gissa.org.za/special-interest-groups/open-source/gis-tender-examples. Joburg, DRDLR and Umgungundlovu do not meet the above criteria and should hence be rewritten and re-issued. eMalahleni is an example of a much better-constructed tender that on the face of it provides a level playing field.

Or these two atrocious examples:

• "COR 8/1/1/09: Molemole Municipality is hereby inviting quotations from service providers to render a service of supply, delivery, installation and configuration of ArcGIS licenses. Specifications: 1. Level 2 Term License; 5 user Pack; 2. Level 1 Term License; 10 User Pack; 3. Service Credits; 1 Block (1000 Credits); 4. Map (1); Drone 2 Map for ArcGIS Term License. Please note that this quotation was published late.”
• Mandeni: "13/19/20: Quotations are hereby invited from suitably service providers to supply and deliver of Software Licenses as per the following specification: • ... • ESRI Software maintenance renewal End User No: 307570"

Here is an attempt at a completely open, functionality-based tender specification that you can pick and choose from as a template (this is a wiki, feel free to contribute!)

Competition

"The Competition Commission is a statutory body constituted in terms of the Competition Act, No 89 of 1998 by the Government of South Africa empowered to investigate, control and evaluate restrictive business practices, abuse of dominant positions ..."

There is a case to be made at the Competition Commission to investigate certain vendors' monopolistic practices, which not only flout Government policy but deny opportunity to other businesses, not only to FOSS providers but also other proprietary vendors.

What is Government expenditure on GIS software?

Let's get an idea how big the problem is and how much could be saved by migrating to FOSS

We should request full records of expenditure on ESRI, Hexagon, Oracle and other proprietary vendors by Government for, say the last ten years, with a full breakdown of entities involved and if they don’t comply, to make a PAIA request. Who would we ask? SITA? Treasury? OpenUp? amaBhungane?

Human Resources

SAGC

Universities are partly to blame for the expectation their graduates have of using specific software in the workplace. Most universities acquiesce to vendors' 'education discounts' which are really just entry drugs. Or they use (the vendors') fake arguments that 'we train in this software because that's what the market uses'. University is a place to learn and apply theory and should be equipping all graduates with FOSS tools (as the University of Pretoria does so well). This empowers them completely. If they happen to be employed by a company that uses proprietary tools, it's a simple matter of learning a different interface.

Training

test the market create demand

Support and maintenance

test the market create demand

https://www.qgis.org/en/site/forusers/commercial_support.html

https://www.osgeo.org/service-providers

Are there functional or suitable FOSS alternatives to proprietary software?

list each product and FOSS alternatives

OSGeo

For a comparison of GIS, Servers read http://www.digital-geography.com/arcgis-server-vs-open-source-gis-solutions

For a comparison of GIS Desktop solutions https://gisgeography.com/qgis-arcgis-differences.

Benefits / Pros of FOSS GIS

- starting company - no user or seat limit

1. Immediate creation of many jobs and upskilling people who are involved with transferring and improving data from proprietary software to opensource. 2. Immediate savings in foreign exchange and the competitiveness of our industry. 3. Creation of local skills and products using opensource that can form part of expanding the world-wide opensource offerings as well as offering professional services locally and offshore. 4. Much better information availability from all levels of Government to the citizens that generates optimism on future competence related to the so-called 4th industrial revolution which has seen a lot of talk but very little action. 5. Making much more information available on the web using FOSS and Integration of town and country GIS data with Google Maps etc to have much better information available for Tourists. [Arthur Taute]

Working for Water MIS:

Cons / Disadvantages of FOSS GIS

- can't think of any

Actions to Take

• if you are in government, try to educate managers and SCM practitioners
  • with information in this document
  • with this sample tender specification
• all, share dubious tenders and RfQs on the OSGeo Africa list as soon as you come across them
• submit objections to the issuing entity
• submit compliants to SITA, Treasury and the Auditor General
• submit complaints to the CSI
• get QGIS into your child's school or a school nearby
• get FOSS GIS into your university
• Write to the National Treasury (Chief Procurement Office) and CFO of the Department bringing this to their attention and seeking clarity whether this is inline with Section 217 of the Constitution
• work with an investigateive journalist like Scorpio or amaBhungane, or and NGO like OpenUp, to investigate and expose
• lodge a complaint with the Competition Commission

A suggestion from a list member: This is strictly about getting Open Source to have a fair look-in, and not about the specifics of any tender or purchase. My support is for leveling the playing field, and for exposing/wiping out badly constructed Bid specifications.

• Perhaps you could do some initial research, maybe applicable laws or policy guidance documents.
• Perhaps listers working in government can point to some relevant documents on this matter.
• Then find some legal-minded person to cost out a strategy, and of course make a 'feasibility of success' statement.
• After that, we could have a worthwhile direction.
• With a feasible direction we could try some crowd funding - or maybe even just have to make some representation (through SAGI??) to some government body with the power to change things.
• providing we have a direction and a crowd funding platform, I will pledge R1000 to further this cause.

References

Inputs from the community, particularly from the OSGeo Africa mailing list. Threads like this one or this one.

FOSS policy in South Africa

• http://gissa.org.za/special-interest-groups/open-source
• http://gissa.org.za/special-interest-groups/open-source/foss-documents
• http://www.dpsa.gov.za/dpsa2g/egov_documents.asp

Some International FOSS policy and case studies

• Case_Studies#Open_Source_Policies