Predictive Font Glyph Delivery for Bandwidth Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing font delivery systems incur bandwidth, storage, and time costs due to the need for multiple requests and redundant font file transfers when users enter characters incrementally, as they require sending subsets of glyphs for each character entered, leading to inefficiencies in content rendering, especially for non-Latin scripts which occupy significant storage space.
Innovation Solution
A method that predicts subsequent characters based on entered characters using a prediction algorithm and language models, generating a glyph set containing only the necessary glyphs, which is then sent in a single response, reducing the need for subsequent requests and optimizing font file size by using a cache for superset glyphs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If a font file containing only a subset of glyphs is sent in response to each character entry, then the font file size is reduced, but multiple requests and transfers are required incurring bandwidth, storage and time costs
Solution Approach 1:
The server performs preliminary action by predicting the set of characters the user will need before the user actually requests them. When the user enters the first character, the server uses a language model to predict all subsequent characters that will be needed, and sends a font file containing all predicted glyphs in advance. This eliminates the need for multiple sequential requests and reduces the total time cost of font delivery.
Solution Approach 2:
The system uses feedback from the language model's prediction accuracy to optimize future font deliveries. By monitoring which predicted characters were actually used and which were not, the system can refine its prediction strategy and adjust the set of glyphs included in font files, improving the balance between file size and completeness over time.
2Reliability
If a font file containing all glyphs is sent, then all characters are available immediately, but the font file occupies significant storage space
Solution Approach 1:
The system extracts only the necessary glyphs from the complete font file based on prediction. Instead of sending all glyphs, the language model analyzes the context and user input to identify and extract only the specific characters that are likely to be needed, creating a customized subset font file that is both small in size and complete for the user's needs.
Solution Approach 2:
The font file is customized with local quality by including different sets of glyphs for different users based on their specific needs. Each user receives a font file tailored to their predicted character requirements rather than a universal font file containing all possible characters, optimizing storage efficiency while maintaining reliability for each user's specific context.
3Loss of energy
If the glyph set is dynamically adjusted based on prediction, then bandwidth and storage costs are reduced, but the complexity of the system increases
Solution Approach 1:
The language model acts as an intermediary between the user's character input and the font file generation process. It mediates by analyzing the input characters and predicting the set of glyphs that should be included, translating user input into optimized font file configurations without requiring complex custom processing for each request.
Solution Approach 2:
The system changes parameters of the font file (which specific glyphs are included) based on the prediction results. By dynamically adjusting the glyph set parameters according to the predicted character needs, the system optimizes bandwidth and storage usage while using well-established font file formats and structures, avoiding unnecessary complexity.
Data Source
AI summary
A system 20 and computer-implemented method is disclosed which involves receiving at a server 24 a live query that is being entered at a client location 22 in relation to creating an electronic document, the query comprising plural characters that have been entered. The system 20 and method uses the query to predict characters that might follow the plural characters that have been entered at the client location. Next, a glyph set is obtained which includes the plural characters that have been entered and the predicted characters in a font, the glyph set being absent of other glyphs of the font. The obtained glyph set is then sent to the client location for use in creating the electronic document.