Latitude® Change List history:
Latitude® is the Enterprise-Grade Project and Task Management Software System utilizing a task-based simple AI implementation for increasing system efficiency through Process-Interaction and Self-Evaluation (PISE™). Latitude® is a flexible and powerful system management solution empowering best practices for information technology services through “Task Enablement” using the PISE™ interactive algorithm. Latitude® task management software enables software tasks to be more organized, productive, and effective through intuitive automated tracking, prioritization, and self-evaluation to increase overall system efficiency and effectiveness by 75% to 200%.
Monautocracy® – pre-alpha early limited release v0.0.13
Task management methodology assigned through a central routing authority. Tasks are assigned priority and resources based on a pre-determined list of arbitrary constraints handled by the central routing mechanism. Individual tasks are given limited determination in regards to allocation of available system resources assigned to them via a stub loader activated by ‘envoy’ sub processes of the central routing authority. Rudimentary self-improvement algorithms are assigned to all system and task functions based on the centrally assigned priority and available resources.
Bug tracker ticket #193.
Summary: Regardless of the arbitrary constraints, the AI based decision making of the central authority seems to be mis-allocating system resources in favor of the efficiency of the central mechanism itself. The self-improvement mechanisms make this problem worse in that the ‘self-evaluation’ methodologies ultimately favor the central router which continues to assign itself more system resources in that it is inevitably self-designated as the ‘most important’ function in the system. What ‘self-improvement’ was noticeable seemed to focus on improving the ability of the central authority to horde those resources rather than improve the overall system performance. Individual tasks are then assigned less and less system resources from the diminishing quantity that remains. As a result, the self-evaluation mechanisms of those individual tasks suffer as a consequence resulting in less individual task performance and virtually no individual task self-improvement.
Oligarcus® pre-alpha early limited release v0.1.03-c
Alterations were made in the task management scenario to instead assign task priorities from a pool of authoritative ‘delegate’ sub-routines, each with an individual arbitrary list of priorities and parameters in specific categories. No significant changes to the self-improvement algorithm was implemented at this time at the individual task level as this functionality was assumed to suffer primarily from inordinate distribution of system resources in the previous alpha release.
Bug tracker ticket #271.
Summary: Again, regardless of the arbitrary constraints assigned to each ‘delegate’, the AI based decision making is still skewing over time to favor the delegate processes. In some instances, even despite of the over-allocation to these components, the ‘competition’ between the delegate applications for these resources further lowers overall system performance through repeated inter-process challenges that one developer dubbed ‘filabusting’. The self-improvement mechanisms again seem to lie primarily at the administrative level and again focus more on improving their ability to ‘horde’ resources to the benefit of the individual ‘delegate’ apps. This in turn degrades overall performance. And again, the end result for the individual tasks was not dissimilar to that of the prior alpha release.
Democra® alpha limited release v0.2.185
Problems with misallocation of resources in pre-alpha releases is now being addressed by increased prioritization at the task level. Individual tasks now include an evaluation algorithm and each contribute individual input toward the ultimate assignment of overall system usage. The central mechanism is now serving solely in an administrative capacity to evaluate and apply the combined prioritization.
+ New Feature: the individual task evaluation algorithm will now be categorized under the name PISE™ for Process-Interaction & Self-Evaluation at the individual task level.
Bug tracker ticket #1393
Summary: When assigned limited scope tasks, the system seemed to perform reasonably well. But upon assigning more complex, multi-function task scenarios, the individual prioritization of system-wide task priority does not take into account enough parameters as to the importance of those tasks actually needing system resources for priority functions. The result is that even critical system functions are sacrificed to the collective ‘whims’ when the individual task evaluations are combined and evaluated. Individual task self-improvement, nearly non-existent in prior releases, does show promise and shows considerable improvement in overall system function when not confronted with the aforementioned critical task allocation issues. But the elimination of vital tasks from resource allocation ultimately results in complete system failure when vital system processes are no longer able to execute.
Republi-Bureacra® beta release v 0.3.3-b
Delegate mechanisms re-introduced to handle a more defined prioritization of individual task evaluation results. Added functionality to the delegate prioritization methodologies to auto-generate task ‘agencies’ to manage the distribution of system resources based on final priority schemas implemented following a system wide task-evaluation data census. Inter-management guidelines are now set in place to allow delegate processes to self-monitor one another in addition to task population as a whole in order to assure that no one delegate processes consumes unreasonable quantities of system resources.
Bug tracker ticket #2743
Summary: Overall prioritization performance is greatly improved… initially. The self-improvement process is working excellently at both the individual task level and the delegation/task-management level. However, over subsequent generations of self-improvement, the delegate code evolution seems to be prioritizing the generation of more and larger agency stub applications to handle allocation and distribution of resources until eventually almost all system resources are consumed by both the agencies and the individual tasks when trying to comply with the plethora of parameters and restrictions put in place by the various agency stubs.
Communas® v0.4 beta release patch v0.4.12-4
Delegate sub-routines now maintain tasks policies through a policing authority similar to the Democra® alpha release. Task evaluation on the individual level is again pooled, but the individual algorithms modified with prioritization schemes favoring an arbitrary list weighted to focus primarily upon improving overall system performance and maintaining the pre-determined task priorities. Delegate ‘policing’ was also modified and strengthened to maintain and enforce these pre-determined priority scenarios.
Bug tracker ticket #3288
Summary: Decreased performance witnessed from the start. Self-improvement mechanisms of individual processes seems to be directly effected (stalled) by the prioritization of pre-ordained system-favored outcomes. While there was little or no individual task improvement, the delegate process self-improvement continues to favor maintaining the arbitrary list of rules. This ultimately produced a result not unlike those seen in the Oligarcus pre-alpha version. Overall, resources are seriously mis-allocated and despite trying various arbitrary priority scenarios, there always seems to be a severe shortage of system resources upon allocating them to the task population.
MonoCorp® beta evaluation release v0.5.13
The focus of task evaluation was again returned to the individual task level, but the means of rating other tasks’ priorities is now based on direct interaction between the individual processes themselves. (tasks that have no interaction with other tasks make no evaluations on these tasks) Tasks compete for and exchange system resources based on the value they provide to the system as a whole as evaluated through their interaction with other tasks within the system. System allocation of CPU usage is maintained through ‘CPU credits’ allocated based on total available resources.
+ New Feature: DeFedCit™ allocation - As part of the new CPU credit scenario, we have enabled the delegate level sub-routines to make predictions (called ‘Next Evaluation Supply’ or NES requirements) of future requirements as determined by pre-defined guidelines. By focusing on Key-NES predictions, this should properly allocate system resources over time as well as in the immediate moment. To help facilitate this, tasks with predictable future priorities may now exchange some of their allocated resources for the purpose of gaining increased priority in future allocations from the delegating authority process(es).
Tasks can now also solicit other peer and non-peer tasks with CPU credit surpluses based on their own evaluated priority and performance. These credits may then be requested later by the ‘loaning’ tasks if needed and the amount will be scaled appropriately based on the gross CPU balance of the task that solicited them comparative to gross CPU balance at the time of the initial exchange.
Bug tracker ticket #6374
Summary: Overall initial performance greatly improved. Cross evaluation of task importance is highly effective in that those tasks most vital to system performance gain the best evaluations from those peer tasks they assist and ultimately end up with the appropriate level of available resources. However, the ability of tasks to exchange resources on the delegate level for future favorable allocations seems to be backfiring. Those tasks receiving the most system credits based on their peer-evaluated exchanges are ultimately able to purchase the most future-priority from their respective delegate regulators until they end up controlling entire sections of system task prioritization. This mis-allocation can even spill over to effect other similar task categories. Ultimately a small handful of favorably peer-evaluated tasks end up controlling the vast majority of system CPU credit distribution through a process of ever-increasing delegate priority ‘grafting’. Furthermore, the ‘prediction’ scheme at the delegate level seems to generate an ever-increasing amount of ‘future credits’ by fiat. Over multiple cycles this ends up leaving lower-evaluated processes waiting in long queues even to gain access to those system resources they have sufficient CPU credits to utilize.
Mixedeco® v0.7.7 pre-release
The focus of task evaluation strategy so effective in the prior release was maintained at the individual task level through peer rating. The CPU credit system was strengthened but most of the interaction at the delegate process level was removed and allocation is now based entirely upon currently available resources ONLY. (individual processes are now required to making their own predictions on future required resources and are left to plan accordingly for their predicted needs) The ability to ‘pre-purchase’ favorable priority was removed and delegate processes altered to primarily regulate the exchange of CPU credits among the individual processes to assure fair distribution of resources based on pre-determined (regulatory) guidelines.
+ New Feature: Introduction of InnoVA™ to the self-improvement AI. InnoVA™, which was partially implemented in patch 0.6.483, allows individual tasks to recognize new system task requirements not necessarily directly related to their own existing task functions and not yet implemented within the system. Those newly spawned InnoVA™ tasks deemed favorable through peer evaluation can then be incorporated as a new system function running autonomous from the direct management of the task who’s AI initiated it. Occasionally, multiple InnoVA™ processes may be spawned to address identical system needs, but peer-evaluation should increase the evolution of the self-evaluation process so the most efficient methods gain priority over time.
+ Alteration: After considerable review we have removed the pre-allocation capability of the DeFedCit™ enhancement but have left the peer-to-peer allocation of CPU credit surpluses intact as it increased performance in the prior release by allowing for inter-process sharing of otherwise dormant resource allocation credits. Additional parameters have been added to the individual tasks’ AI equations to assist in the proper ‘individual negotiation of value’ (IN-V) of surplus CPU credits. This feature will be renamed to IN-Vest™ prior to the final market release.
Bug tracker ticket #11481
Submitted by beta-tester Jeff Thompson (jthom@*******.biz)
Great package! I think you guys almost have this licked. I was able to improve my system performance by 50-75%!!! One suggestion though based on my own tinkering. After running the Mixedeco package through a few generations of self-improvement cycles I began to notice some tasks gaining increased priority. It wasn’t incredibly out of proportion, but it was enough to grab my attention. Upon further examination I realized that the vast majority of these processes were the results of the InnoVA task creation. There were a smattering of others also that, when combined with the InnoVA tasks gave me a theory.
The similarity between all of the increased priority tasks was that they were universally unaffected by the restrictions I pre-outlined within the arbitrary rules at the delegatory level. Based on this assumption I re-initialized the system with only the bare minimum of regulatory guidelines necessary to direct the prioritization and self-improvement methods toward the goals I desired. It took some tweaking, but after just a few cycles I am easily seeing 200-300% improvements compared to the initial 50-75%!!!
My suggestion to anyone using the management suite would be to keep your regulations to as few as possible!! The self-improvement and self-evaluation schemes seem to be far more capable of generating desirable results with less inhibitors in place. Keep up the good work!
Latitude® v1.01 final release
Based on beta-tester recommendations, the regulatory layer’s impact was greatly reduced for the final release. Most additional patches at this time involve improvements in the user interface and only minor tweaks to the existing mechanisms. Latitude® is scheduled for release early next spring complete with PISE™ and IN-Vest™ technologies!